Saturday, July 31, 2010

Still More Size

Some investigation shows the following:

  1. Because of the way "Level 1" works the first 8 AP points spent on dealing damage result in more damage than each 8 AP thereafter. For PEN STR (a short range punch-like attack that does PEN damage) the numbers test at 13 PEN for Level 1 and +7 PEN for each 8 AP thereafter. When applying that to Armed Size the first 8 AP level of Size would do, maybe, 6-9 PEN and then add +4 PEN thereafter. Doing this helped Armed Size (what I'm testing right now) greatly. But not greatly enough.
  2. The FULL ARMOR + Armed Size wins about 75% vs. 40% for the others. This includes the partial armor/partial DP build. This means that "Armor isn't simply too cheap"--the build with a little armor and a little DP gets beaten like everything else. It's the big investment in armor that really pays off.
  3. As suggested previously, the 30% offense/70% defense builds of the normalized herds seems ill-equipped to deal with heavy armor and extra DP but does well against everything else.
  4. Doubling the Size DP for the FULL DP character did help--but not enough.
E's Build
Some discussion led us to the conclusion that a character with "half their points in Defense" and "half their points in size" as a Low Damage build was probably not a "common build" (even if, when that defense was FULL ARMOR it was shaping up to be very effective). We decided that a more likely build was: 8 AP in size (that's, roughly, 4 AP in damage and 4 AP in DP) followed by a 4 AP weapon and 4 AP of some defense (such as Armor, Force Field, more DP, etc.)

This "common build" was 50% offense and 50% defense. I tested it with FULL ARMOR against the 16 AP Normalized Herd and it won a (much better) 66%. This is not good but it is better than 75% by a large margin: it's potentially just a "good build."

Next Steps
I plan to test that build: 8 AP Size, 4 AP in a Sword, and 4 AP in either FULL ARMOR, ARMOR+DP, FORCE FIELD, or FULL DP. If the spread of these is around 50% then we're golden.

Thoughts on FULL ARMOR
Clearly putting half your AP into armor and nothing into DP is a reasonable strategy from a statistical perspective. If that's the case against virtually any range of opponents then that's a problem: it's not good for gaming for there to be a single best-always defense. There are some possibilities and some thoughts.

  • Poison / Ignore Armor. There are attacks (Poison Gas Cloud) that ignore armor. These mess up FULL ARMOR badly as it lacks the DP to defend itself. On the other hand, FULL ARMOR + Poison Gas Cloud does better against traditional attacks than FULL DP + Poison Gas Cloud so as long as you aren't facing Poison Gas Cloud FA is still a good way to go. Worse: these attacks are almost unquestionably exotic. Almost no game I have ever played in features Resisted Attacks as often as normal ones.
  • FULL ARMOR Wasn't That Good In The 50/50 Herd Test. The current "cost" for armor is based on its tested performance in the original 50-50 herds. Those were fast, brutal battles where the heavy attacks would tend to overwhelm most defenses anyway--but the fact remains that in some battles it's simply not that good. The current 30%-offense Normalized herds tend to overvalue heavy investments in armor in the way any given game might not.
  • Egg-shell Syndrome. We think there will be a "blow-through" problem for FULL ARMOR PCs: in any battle where the damage being dealt can penetrate the armor reliably they won't have a cushion and will go down quickly. The question that remains is: how much damage is needed to do this and is it even true that when that much damage is being done FULL ARMOR will actually fare worse than, say FULL DP or Mixed (in fact Mixed DP&ARMOR may be the big loser since it doesn't get the extra DP that Full DP does and doesn't get the benefit of all that armor either). Finally, if FULL ARMOR even does fare worse against heavier attacks, how much heavier must these be? Who fires them? Are there characters that are 75%-offense? Do "boss characters" throw those attacks heavily? Does an easily purchased "charge-up attack" qualify? We don't know yet.
What Could We Do?
Right now 4 AP buys 4/10 Armor--the equivalent of chain mail. The fact that 8 AP ("one level") can buy a broadsword + chain mail has "beautiful" implications for our costing system (even if, in fact, most Fantasy characters would not likely be paying AP for swords and armor). I really, really want to keep these numbers and, in fact, all our testing has been based on this.

So what could we do?

  1. Not sell "raw armor." We could sell Armor + DP in all/almost all cases. In fact that's not far off from what we're doing now. This makes it hard to simply get FULL ARMOR combined with 50%-offense (Note, however, that the FULL ARMOR + 25% DP + 25% Offense character did very well against the Normalized Herd and this would still be quite possible). The problem here is with gear. I refuse to sell a "bullet proof vest" that "comes with" DP. It doesn't make much sense.
  2. Give out less armor after Level 1. We could reverse the Attack Theory and give fewer armor points as more armor is purchased. This would reduce armor's value at the upper point levels. Although this might require some re-testing it probably wouldn't require much. The problem is that every ability that grants armor would have to be couched in those terms and I'm not sure how well that would work: There are likely to be some abilities that grant armor that just have "one level."
  3. Increase the amount of DP/Force Field/Etc. You get "per point." This sounds like a no-brainer but note that FULL Armor + DP in the Size Test is faring better than Force Field + DP right now or even HALF ARMOR + DP (Mixed). Giving FULL ARMOR even more DP won't necessarily fix that. Another big issue too is that this will change a lot of the testing we did (although, again, maybe not by too much).
Thoughts So Far
I'm going to test the spread of defenses as stated above and see what I think. Then we'll do some testing to see if slightly higher attacks than 30% can actually exploit the low DP that FULL Armor grants. If it can then I think we're golden right now.


Thursday, July 29, 2010

Size Testing

I did some preliminary testing of the Size equations from last post. What I discovered was disturbing. It's going to be hard to go into in a short blog post--but what it amounts to is this:

1. In the current "Normalized Herds" there are 16 Opponent Characters and 4 Test Characters. Each of the 16 Opponents is 30% Offense and are armed with sword, gun, punch, and blaster (with various combinations of defense), 70% Defense (calibrated for a 3-4 Round fight). The Test Characters are 50% Offense, 50% Defense and will have whatever weapon is being tested. One Test character has FULL armor, one has FULL DP, one has half armor and half DP and one has a Force Field.

2. When I tested the Size Equation with Armed Size (1 STR = 5 BLD = 5 DP, character does PEN damage) I got, for the 16 AP Herd (being 8 AP investment in Size and 8 AP investment in defenses) a 50% battle-win for all four characters at 4 STR, 20 BLD, 20 DP.

That would be good: except that the issue is that our current pricing gives 20 DP by itself for 8 AP. This would effectively mean that the offensive power (STR and BLD) was free. As I was doing "Armed STR" the damage dealt by the character was expensive PEN damage ... so making it "free" wasn't good.

3. Worse: the spread across 4 characters sucked. The FULL ARMOR guy was winning like 75% of his battles while everyone else was around 40%. The FULL DP guy was only winning like 30% of his fights. This spread was technically 50% but the numbers aren't good (well, they're good if you're the FULL ARMOR guy but otherwise Size is a sucker's game).

Now, I hadn't factored in the +'s TBH for being big--or the extra reach--but at the 8 and 16 AP levels these aren't so much. Not enough to make a real difference anyway.

What Was Going On?
Well, first things first: we determined that  for the FULL ARMOR guy the fights were taking like 6.5 Rounds which was damn long. We also figured out that the mix was ... problematic. When we were putting 8 AP of Size into a test character the character had: 8 AP of Defense, 4 AP of DP (also a defense) and 4 AP of Damage. This made the character a "Low Damage" character.

What Does "Low Damage" Imply?
Well, for starters it means that characters who elect to have at most 1/4th of their AP in offense usually get some kind of special deal for it (such as being fast for a lower cost). In our case the characters were coming out as LD characters but weren't getting anything (well, technically they were getting attacks for "free" but that made things even worse).

Secondly, when they were facing a herd of 70% Defense characters they were seriously under-gunned compared to the 50% characters I'd been testing with.

What we determined was that the herd itself was not a good test for a power that split offense and defense.

Final Issue
The final issue is that we've determined that characters without armor should probably get extra DP when they buy DP--like maybe double. Our builds don't come with armor and we're not currently testing them with double DP. Size doesn't imply armor so if we were to invest a full 16 AP in Size (which, at this test rate would give +8 STR, +40 BLD, +40 DP) the character would have no armor but would be approximately 50% of his AP in offense.

I'm not certain what to do with that. For a 16 AP character do I test 12 in Size (6 in Offense, 6 in Defense, approx) and then put 4 in a second defense? Or do I do 14 in Size and 2 in defense? Which is "more likely" as a PC build? Maybe both?


Wednesday, July 28, 2010

Size (Does Matter)

I'm still testing toxins and pincers: The scripts are taking a long time to run (maybe I need to reboot?) and I'm concerned about data quality for some of the toxin results (two unlike--but similar toxins came out identically powered for the 32 AP test and that ... makes me wonder).

But in other news, I'm working on a plan for Size.

In JAGS the way you play a grizzly bear or a giant or a godzilla (correctly called a Kiaju--the genus of those things) is by buying Size to make you bigger. This is in addition to getting claws and teeth, a huge club, or nuclear breath. Size is like "Strength" except:

  • You MUST have at least 1 STR per 5 BLD.
  • You MUST have at least +1 DP per 1 BLD.
  • You get bonuses To-Be-Hit (which may vary depending on if the attack is Ranged or HTH) as you get bigger
  • You get +1 CON above some level (roughly the weight of a tiger)
  • You get more 'Stride' in the form of bonuses to ground movement
  • Your bio-weapons (or just fists) get more Reach
  • You may get +1 Large Weapon Bonuses to hit things that are smaller (there are some optional rules around this)
This is all pretty complex. Let's see a simple example:

Example: I am creating a creature that weighs 300lbs and I plan to get that by buying Size (aside: there are other ways to get BLD and such, but Size is the most complete package). I start with a base character that weighs 150lbs. I need 150lbs of Size. So here's what that's like: (1) It will require 10 BLD (150 lbs). That means the package MUST also come with +2 STR and +10 DP. If I can just "buy Size by the BLD-pt" then I get +10 BLD and the package must include the other stats. (2) I can see on a chart that this size gives me +1 to be hit at Range and gives me a modest stride increase. (3) There is no Reach or CON increase and my humanoid size, if I am humanoid, is estimated at 6'6".
Okay, so how do you "sell Size"?

Selling Size
The JAGS convention is that Traits usually cost either 4 or 8 AP per level (it's not a rule but it's a guideline) and if you are buying STR/BLD you must specify whether the character will be Armed or Unarmed. The question is: "Does the character commonly use a PEN-damage HTH attack?" If yes, you pay the Armed cost and if not you get the Unarmed Cost.

So far, so good.

The problems here are as follows:

  • My costing-methodology works very well for attacks but we have never tried it out with the mix of attacks (STR/BLD) and Defenses (DP). I'm testing now to see what happens and by the end of this post maybe I'll have some stats for the preliminary run.
  • None of this tells us how to cost for +'s to be hit and CON (both of these are, costed correctly, a percentage of your TOTAL AP points). 
  • As we would like to sell Size "by the BLD-pt" that breaks the 4/8 AP rule pretty immediately.
Decisions About Unarmed Size
I have to determine what my "rule" will be for Unarmed Size. As stated above, the requirements for STR and DP are minimums. A creature must be strong enough to "lift itself" (the requirement that each +5 BLD have a corresponding +1 STR) and it must be "appropriately tough" for its size (the requirement that each pt of BLD have a corresponding Damage Point). However, those are minimums.

Creatures with Unarmed Size (like giant superheroes or, like, horses) will be either cheaper, stronger, or have more DP than their equally sized Armed brethren. I need to decide what I'm going to do about that--and it's probably going to be some combination of the three. Testing shows that Unarmed characters can be cheaper and stronger (although they must, of course, keep the BLD the same since that's what you are buying).

What About CON, +'s TBH, Stride, and Reach?
It looks like those will come off a table and you'll index your size (BLD) and Total AP and see what the cost is for these "extras." As bonuses To-Be-Hit are bad for you, this may well wind up be negative points (you get points back)--or maybe we can just give you more DP (my preferred position) and that's what you'll be looking up.

I'm not sure yet.

End Note
I'm still doing some testing--but one consequence of an ability that adds DP as well as Damage is that the ALL ARMOR guy comes out way, way better than the others for reasons I'm not entirely clear on (well, I'm mostly clear--the ALL DP guy gets an additional multiplier for his DP that does not apply to the Size DP granted). I'm still running the scripts for the 16, 32, and 64 AP levels.


Tuesday, July 27, 2010

Success Points

I'm back from comic con! Today I'll be testing toxins, spines (like a porcupine's quills--it attacks you when you attack them!), and pincer (a grab/worry style attack).

I got a series of questions from a reader to work on. I'm starting here with Success Points. But before that ... a quick note on Disease.

We actually have had some notes on disease at various places in the rules. The problem with disease is that (a) it's seriously un-fun but moreover (b) it serves some very strange purposes in RPG-play. I love the nasty colorful diseases from Warhammer. If we do a dark fantasy game it'll probably have some notes. Diseases can be a sort of slow wasting that spur PCs to action (we must save the prince by getting a healer!) or they can be simple 'realistic' hazards of injury or environment (don't go down in the sewers, you'll get skin-rot!). If a PC is diseased then usually it means either (a) it's part of their character and is long term of (b) it's simply an obstacle that has to be overcome in the short term. I suspect dying of disease would be very unsatisfying.

Finally there's the carrier issue: in a post-apocalyptic wasteland it might be cool to have a disease carrier character who would spread doom on towns he crossed (or, you know, maybe those who spilled his blood?).

In any event the issue is two fold: the first is a question of infection. The deadliness of a disease is not proportional to its infectiousness. In fact, the really deadly diseases are probably harder to catch than the other way around. JAGS Disease rules will need a way to handle infection apart from the actual effects. This would likely be a CON roll modified by the disease itself and circumstance (if you bathe in it, roll at -4).

The effects themselves (once you have it) would indeed be a Resisted Attack with long-term effects (most diseases will not kill you during combat). The rules for curing would likely be a Resisted Attack roll vs. some portion of the initial Intensity based on the Result (i.e. a Catastrophic Result might be 3x the initial Intensity, while a Minor Result might be 1/4).

There is probably a note for each level at how often the PC has to re-roll their defenses hoping to get better but possibly getting worse. These are complex rules and would need at least a page or two in order to be covered. They're low priority for the next release.

Success Points
Success Points are fairly new to JAGS and they exist because of a couple of specific needs. These are:

  • A way to measure the progress of a Drama Roll. A Drama Roll is a series of rolls (usually 3), where the amount each roll is made by is tracked against a Target Number. So, for example, if your safe-cracker is working on the safe (and gets one roll every 20 min) and they need a total of 20 SPs to crack it, over a given time they can likely eventually crack the safe--but each 20 min it's likely something happens (roll for a guard to come by?). This creates DRAMA!
  • A 'meta-currency' in the game which can be expended during play to make the character more effective. For example, a character with the trait Loser has Success Points and can loan them to other players. The PC, as seen by others, is a bit of a hapless dork--but his or her friends get "buffs" from being around them. This is sort of like Xander in Buffy the Vampire Slayer.
These are two different things and they (confusingly) share the same name.

What Are We Going To Do?
Well, it's very possible that we'll change the name of one of the two things. I'm open to suggestions. However, as the two are somewhat inter-related let's discuss in more detail how they should work.

Do You Get SPs From a Drama?
If the safe-cracker in the example above needs 20 SPs to crack the safe and generates 30 does the character keep the other 10? No. You don't. SPs generated for a drama are lost when the drama ends. That, however, is not the whole story: sometimes the point of a Drama is to generate SPs. Huh? Let's take a look:

Trait: Con Man. The character can run "confidence games" to generate money and SPs which they can later use. In this case the con itself is a drama: the PC must beat a Target Number with three rolls to successfully pull off a Con against a mark. If they succeed they get some money and they get 2 SPs! A Con Man has a "battery" of 4 SPs per level.
What does the above mean? Let's break it down

  1. The trait itself will be a Generic Archetype Trait and will cost 4 or 8 AP per level. It grants a "special ability" to generate SPs through Con Games. NOTE: you do NOT need Con Man to run a con game. Anyone can do that. You don't even, necessarily, need Con Artist skill although without it you will be hugely disadvantaged since the GM can't assume you get things correct when you don't specify them. What the Trait does is grant the PC the ability to generate SPs that they can use for other things (possibly "whatever roll they want to improve").
  2. The rules will sketch out a general drama for a con game. Any PC with the right skills can do it. Again, you don't need the Trait to try to con someone. What changes if you do have the trait is that you get, in addition to money, SPs for it.
  3. When a character with the Trait does the drama and succeeds all the SPs generated by the Con Artist skill rolls will go away (poof!) but if successful the character will then GET 2 SPs to mark on their sheet. These can then be used for other rolls or whatever.
  4. We are working on the concept of a "battery." This limits how many SPs the character can walk around with (4 per level) and answers the question: "how many SPs does an NPC with the Con Man Trait have?" NPCS are generally assumed to be either half or fully loaded.
What Can You Use SPs For?
This is tricker than it sounds--we do have some rules for this in the book ... but ...
  1. The general rule is that you can only spend SPs if you made the roll. The SPs just make it better. This is done because we want your roll to be important in the game--moreso than SPs.
  2. Can you "buff" someone else's roll? Often yes. We expect many SP-pools to be able to be lent to other characters. This is a good way of handling certain meta-game things like a commander who improves his men's abilities.
  3. What about restricting SPs to certain kinds of rolls? We plan on it. Imagine a magical sword that gives 3 SPs for dealing more damage or getting around a block. That's offensive. What about a magical shield that gives 4 SPs for making a block with it more likely to work or improving what a CON roll was made by. Those are defensive. There are several other possibilities (a lab that generates SPs for science rolls?) 
  4. How do SPs effect Resisted Rolls? We need to determine this because those rolls are important and SPs need to have impact. Likely an SP can improve the roll by 1 per pt spent? Something like that?
When Do You Use SPs?
We need to understand how big a deal SPs are by testing in the simulator. This will be especially key for magical treasure. Here are some "use cases"/scenarios that I foresee for SP usage.

  • Improve hit-by to 4+ for PEN damage. The character will only spend SP when the hit is by +0 to +3 and will get it up to +4.
  • Hit Around Armor. The character will spend SP on any roll that hits to hit around plate armor.
  • Get Around Block. The character will spend SP to get around a block. How does this happen when dealing with the defender spending SPs as well? Does it go back and forth? I'm not 100% sure.
  • Reduce Armor Save. Each SP will reduce an Armor Save by 1pt. The character will spend the Armor Save down to an 8- if possible.
  • Make block work. When the block roll is made but not by enough SPs will be spent to shore it up.
  • Improve CON roll. SPs are spent to move a CON roll towards No Effect.
  • Improve Armor Save. The character can spend SPs at +1 to the Armor Save for 1 SP before the roll is made to ensure it will be made. The character will always spend up to a 12- if they have the SPs.
What Else?
There could be SP pools that regenerate during combat (the sword recharges every time it kills an enemy) and so on. We should think on that (what if the shield recharges when it Blocks five attacks?).

Wednesday, July 21, 2010

On Hold For Comic Con

I'll be doing updates when I get back. Right now I'm on the west coast attending Comic Con so I'm not testing anything. A few notes, however.

Toxin DQI
I discovered that some of the testing I'd done yesterday and the Friday before had a code error that made several different toxins all the same thing. I discovered this when I looked at results for a weak one and a powerful one and they were identical. It took me a while to run it down though.

Toxin Plan
Ideally there would be one "damage dealing toxin" and it would just "scale." In fact, we could probably do this--but I'm less comfortable about that than I am about increasing the Intensity Score. On the other hand, if a toxin does X-damage points at each level then, at a certain AP range, no matter what its Intensity score it'll be irrelevant: imagine a toxin that, at Catastrophic Effect level, does 30 damage: if used on an elephant with 100's of DP, the Intensity isn't really all that relevant. It can score Catastrophic every time and the target elephant won't care all that much.

So what I have done is create a list of toxins (right now 3) from light to heavy doing damage at three different ranges. The idea would be that if these plot well I can either have a "make your own rule" that shows how to extrapolate or I can create one toxin and determine how to scale the damage AND Intensity against APs.

Either way, hopefully this is almost the end of this whole toxin thing.

Toxins Aren't Fun
Jeff pointed out over the weekend that Toxins are not (or are rarely) "fun." PCs aren't known for using poisoned weapons and being poisoned is--just--not. fun. I agree. The discussion about why I'm doing this anyway was interesting.

1. Being poisoned creates a certain kind of dynamic tension. The PC is taking recurring damage and their fellows may need to intervene.  This is interesting in some situations.

2. Facing a toxin-using foe (giant scorpions!) creates a different kind of tension: the character really doesn't want to be hit and each individual toxic strike is more dramatic than just another claw-hit.

3. There's a verisimilitude argument: if my alien has poisoned fangs then I want to be able to simulate that. Even if it's not a big part of the game, it's good to have it.

4. There are a lot of effects that are beneficial to the game that are similar to toxins but not identical. Things like sleep-darts, paralysis gas, and mental attacks all use the same or similar rules. These are definitely good for the game (having cops be able to use 'Tasers' that are properly represented in the rules in a way that is (a) balanced and (b) feels 'right' is something almost no other RPG I am aware of does).

So anyway, that's the reason we're spending so much time on this.


Thursday, July 15, 2010

Talking Toxins

One of the hazards of this blog is that I'm going to spend time (DAYS!) talking about stuff that you cannot and (I'll wager) should not give a crap about. In this case: more on toxins.

Today I ran the scripts to test TOXIC CLOUD. Toxic Cloud bypasses all the "draws blood" nonsense and simply hits the target with a Resisted Attack that they must roll against. It's the same thing we had before: it does immediate damage and then continuing damage until they make a CON roll at some negative (usually -1).

Given the JAGS Resisted Attack rules I realized (I saw it yesterday but didn't quite know what to make of it) that there was a problem: the problem is that a given damage-dealing toxin is NOT suitable for all AP Ranges. Let's see what I mean by that:

Average Damage Points vs. Intensity of the Toxin
A check of the 32 AP Herd shows me that the 30%-Attack builds (16 characters) have, on the average, 42 DP. The average for the 4 Weapon-Using members (the 50%-Attack builds that I give a weapon to in order to test it) have, on the average 31 DP.

The Toxic Cloud at the 32 AP level wants a 65 Intensity in order to win 50% of its fights assuming the Toxic Cloud Attack operates at range, gets +2 to hit, and fires twice per Round. According to my calculations that means it has a 16- roll and, if the attacker rolls a 6 or less it'll deal its maximal Catastrophic attack of 24 pts of damage immediately and 8pts of recurring damage. More likely the attacker will roll a 10 and get a MAJOR EFFECT which will do 16 damage followed by 6.

These are all respectable numbers.

But when I outfit the weapon in a "more likely" Breath-Weapon configuration of firing once per Round? It wants an Intensity of 140 (Roll of a 30-) in order to win 45% of the time. In other words, the somewhat meager damage that this particular brand of toxin does will NOT win you a fight (against powerful fists, blades, a .357 equivalent gun, or energy blast) unless it regularly does its maximal effect.

Even worse, the stats for a 'breath weapon' I'd like to have include 1x-per-Round fire and a Cool Down round or even Charge Up (so the Toxic beast can't breathe on the party 1st Round). I know what my simulator will say: probably no amount of Intensity will give that configuration a win.

Not with that toxin.

Other Toxins
Of course I'm free to build any toxin I want. We have a good number of effects to choose from and can integrate even more with a little effort. I can have the toxin simply deal an immediate and fight-winning unconscious/death result. I can have the CRITICAL EFFECT level do a million points of damage. I can play with this however I want.

What we think this means is that there will need to be different schedules of toxin that will be either more or less relevant at differing point levels. The Toxin I have now is good for a fantasy game level: if your fighter gets a really bad CON roll and suffers 24 points of damage he's probably down and maybe out--but it isn't necessarily an instant kill.

On the other hand, if you are building a venomous super-villain or space-marine bad-ass (40K type with, uhm, a super-toxic boomerang gun or something wacky like that) then you could get a "schedule 2" toxin where the critical and catastrophic effects do like 100pts of damage (and Major is much higher too). The simulator should, in fact, balance this for the lower levels as well: the starting Intensity for a 32pt Herd might be more like 20 instead of 65.

However, with this toxin, higher Intensity scores will allow the toxin to be relevant to higher damage classes.

So we'll see. I need to cook some of these up and then test them.


Wednesday, July 14, 2010

More Venom!

Today I spent most of the cycles testing various configurations of en-venomed attacks. Everything from 1x-Per-Round (Stinger) to 2x-Per-Round (Poisoned claws) to long-range (Breathe Mustard Gas). I used the same toxin for everything as a level-set.

Here's what I've learned:

  • Indeed, taking a toxin gas against high-armor/low Damage Point targets wipes them out. However, the high-armor guys using the mustard gas are still fighting people (for the most part) with conventional attacks and do very well there. Armor is still king in that scenario.
  • The "draws blood" requirement is, in fact, pretty rare. Over a 3-4 round fight even having a powerful toxin-carrier attack doesn't help that much. I'm not 100% sure why but the PEN damage values came out about where they are for non-envenomed attacks. I think this is because by the time you are doing enough damage to draw blood 40% of the time you are already doing enough damage to win the fight with your PEN attack--so the toxin on top is "just gravy." I examined the fights in strong debug and determined that they did seem to be working.
  • High DP helps a lot against Mustard Gas (and is less effective against the stinger since the stinger does a lot of PEN damage). The win %'s against the high DP guys were like 1-4% while being 88% against all-armor attackers. This is one of those cases where the "average" is fine but the actual components of the average are not. I'm not yet sure what to do about this either. An attack that is very good against many opponents and useless against others is not gearing up for good gaming ...
What Does This All Mean?
Well, it means that I'm going to have to apply some philosophy in pricing chemical attacks. The simulator proves that against our herd their value can be normalized--but I have to carefully consider the ramifications of cheap toxin attacks and where I'd want to see those exist.


Tuesday, July 13, 2010


One of the more complex attack types is a carrier venom attack. Think of poisoned fangs, a stinger, or a scorpion's tail. In this case the attack itself may do some damage (a scorpion's tail can kill outright even without the toxin) and, in any event, it has to penetrate armor in order to get the poison into the target's bloodstream.

This means in JAGS Terms:

  1. The attack must be PEN and must hit (duh)
  2. The armor must fail an armor save
  3. At least 1pt of damage or more must get through the target's armor.
These three effects mean the attack has "drawn blood." Then the poison can go to work. There can be more than one kind of poison. First let's review the basics:

Toxins (and other Resisted Attacks)
In the expanded JAGS Archetypes rules Resisted Attacks have a new statistic: Intensity. It works like this: when you are hit with an RA you (or the attacker--the rules specify that the Player usually gets to roll) must make a Resistance Save (using the Resisted Roll rules) comparing:
  • The Intensity score + POWER Stat of the RA to your Damage Points (DP) + Resistance Stat
  • The "POWER Stat" of the RA to the Resistance Stat of yours (usually CON, sometimes WIL). For each point of difference you get a +/- 1 depending on whether your score is higher or not. 
Let's see an example:

A character with CON 12 and 14 DP is hit by a Taser with a POWER stat of 11 and 15 Intensity. The Resisted Attack Roll is: Attacking Taser: 15+11 =26 vs. 14+12 = 26 or a 10-. Then, adjusted for the character's stat (CON) it goes to an 11- Resistance Roll since the PC has a CON of 12 vs. the Taser's Power of 11.
Okay? It's not the simplest thing in the world but it (a) gives PCs with a good stat a substantial edge which is important and (b) distinguishes from, for example, a pepper spray that will work on a teenager vs. one that might work on a bear. Both of these are necessary for the rules to work properly.

The Roll And Effects
Once you make the roll there are five levels of effect: Minor, Standard, Major, Critical, or Catastrophic. To get a Catastrophic roll you have to fail by 10+ so, in the example above, it's not possible with a standard roll unless you treat a 20 (1:1296) as a Catastrophic failure no matter what. For things that are balanced you'll usually see Minor to Major results. That's what we're going for.

However: a Major result for a Death Ray is very much worse than a Major Result for Tear Gas. The viciousness of each result level changes based on the attack type.

Got that? So here's how this thing breaks down ...

How Resisted Attacks Get Analyzed
Firstly there is their chance to effect at all. In the case of a blood-toxin (as opposed to a gas or "ray") the general power of the carrier (the stinger) is important: how likely is it to draw blood?

Then there's the matter of POWER and Intensity: against a given character point level we can more or less estimate what DP will be and what their stat will be (we assume a CON of 12 for most characters).

Then we need to know how bad a given result is. Many results are "simple." For example with the Death Ray you just take a Wound's worth of DP and make the appropriate Wound Roll. However for some attacks, such as Paralysis toxin, you can suffer lost REA or Initiative for several Rounds. There are negatives Perception Modifiers (for blinding attacks) and Skill Roll negatives ... and so on. There is also additional damage and in some cases it continues for several rounds.

This is all very complex and trying to figure out what the numbers and effects should be is hugely difficult. Thankfully we have the simulator.

What I'm Testing
I started out by taking a baseline and running a test for "Lance," a medium reach PEN HTH attack that hit only once per Round (the theory being that most toxin attacks like Scorpion Tail or Fangs are once-a-Round types of things anyway). Then I created a Toxin attack using one of the basic damage-dealing toxins. I'm running a set of tests now to see how well they'll compare.

I don't know the results yet--but here's an interesting note: usually the character with maximal armor and only a few DP is, in fact, a superior build against peer opponents (although he crumbles faster against slightly more powerful opponents). In this case, once the damage is sufficient to penetrate with any regularity his values are reversed: he loses more often than a moderately armored character with more DP--he's more susceptible to the venom.

Additional Note
Just like IQ Tests measure "how well you take the test" (even though we'd all like them to measure 'how smart you are') our methodology measures "how good this is in the simulator" vs. how-good-this-is in the game. Want an example? Compare Sleep Dart to NeuroToxin Dart.

In the case of the Sleep Dart this is a very powerful attack in terms of result: the Minor Result Dazes you. Everything else knocks you out. It's what you see in the movies: thuck ... thud. Because the results are so bad the Power and Intensity is quite low: if you are going to take out a peer with a sleep dart you either better get lucky or have a lot of your points in it. On the other hand, sleep darts are good for taking out guards and other less potent characters (this is a feature).

So the simulator measures that perfectly.

Next we go to Neuro-Toxin Dart. It has the same Power and Intensity as the Sleep Dart--but the results go: Minor = Dazed, Everything Else = Dead. In the simulator it costs ... the exact same amount (the simulator stops the fight at Unconscious and doesn't care if the target is down or dead). This is one of those philosophical questions: how much is the assassin dart worth than the sleep dart.

Consider that, for a character with murderous intent, in a 1-on-1 fight, putting a target to sleep is the same as killing them if they have a few extra seconds (in most cases, anyway). However, on the other hand, a group of PCs going up against a cyborg assassin who fires shellfish venom darts is a hell of a lot more scared than going up against a cyborg with a tranq gun: so long as the PCs enjoy an advantage in the fight they are not facing roll-or-die (and that's assuming the tranq gun character even wants to kill them).

So ... how much? I have some ideas but I don't really know the answer yet.


Monday, July 12, 2010

Swimming in JAGS

I periodically get emailed questions about the JAGS rules. I do my best to answer them but I know why one of the Champion Designers (I think it was Aaron Allston) wore a shirt that said "I am not a Champions guru." It turns out that not only is my head full of JAGS rules, it's full of several editions of JAGS rules including the copy-shop bound early versions that you'll never see (I don't think I even have any anymore). It's full of various corrections and revisions of the current rules. It's full of stuff I wanted to put in and didn't. It's full of discussions and arguments I've had with the other people who've helped me with it.

When W.S. asked me if there were swimming rules my only thought was "search the PDF. It's what I'd have to do."

Apparently it turns up NOTHING. So let's take a look.

You Complete The Game
The 1st edition of Gamma World had GM's notes that explained that the Game Master "completed" the game. I love this idea. They meant that the act of making an actual scenario was, in a way, "completing" the game--sort of--even if just a little--like making a new set of rules. Of course I don't think the original Gamma World had any rules for swimming either and that didn't mean no mutant could ever swim. If there needed to be rules the GM would be expected to "complete the game" and come up with some.

Here's What I Did (and Would Expect You To Do)
While I was watching Eclipse with my wife this weekend--about the third time or more that the werewolf with the great abs got stalkerfific with the fang-banging emo-girl I started thinking about swimming in JAGS. How would I do it? the question breaks down into a few parts:

  1. What are the (general) rules for being submerged? What are things like movement rates when swimming and things like that?
  2. What does a character need to have in order to swim?
  3. What differentiates a dog-paddling swimmer from Michael Phelps?
In order to figure this out--and these are not "official rules"--I used the universal ubiquitous data-library: Google. I also just applied some general thinking to get rules for the first question.

Rules For Being Submerged/Swimming
Clearly being in the water is hard for people who don't know what they're doing. Plus, they can drown. Movement rates will be slowed, people are easier to hit, and you can suffocate. Also: water would protect you from incoming strikes and stuff.

  • Unless you are an aquatic creature or have some special trait, you lose your AGI bonus in the water.
  • Unless you have some special traits you must spend 5 REA per Round "swimming" or sink. You can hold position (assuming you can swim) or keep on moving or start moving (which will cost you 8) but unless you want to sink, you pay 5 REA at the start of the Round or down you go.
  • Water will make most strikes inefficient. Let's say it confers a -6 Damage Modifier to all attacks (mostly submerged characters get Coverage of 4 that will either result in a miss (simple) or a -6 DM).
  • Knives and spears and other "aquatic weapons" don't suffer this modifier or any modifier at all.
  • If you start drowning you probably fight at -4 unless you make a WIL roll by 4 or more each Round. Skill levels can negate some of this. You can hold your breath for CON Rounds if doing very little but combat is taxing: if you can't breathe water you will start drowning if completely submerged on the 2nd round of combat (assuming you don't spend 5 REA to keep your head above water).
  • A successful grapple can hold someone under water.
  • After your time is up, at the end of the next Round you are unconscious and dying.
How Fast Do You Move?
I asked Google "How Fast Do People Swim" and got a link that said most people at a leisurely pace go .8km per hour. Google calculator (like, the best thing I've ever seen) says that's .24 yards per second or .75 yards per move action. Let's kick that up: an untrained person swims at 1 yard per second for Walking Endurance. 

This is a great link because it goes further to say that if you are determined--but not an athlete--you go at about 2 miles per hour or, doing the JAGS math, 3 yards per movement action. That's Sprinting Endurance.

Finally, it goes to professional athletes. The numbers come out to a top speed of about 6 yards per second as a top contender: being highly trained about doubles your speed.

How Do You Purchase It?
This is a great question. Clearly the generic Sports Skill could be Professional Swimming. So, hey, that way? Well, no--that's not the whole story. My daughter can swim and she sure doesn't have 2 CP invested in swimming (even at L1). I can swim better and I don't have points invested either. While there is a Swimming Skill (in theory) you don't need it to swim a little: that doesn't make sense.

Why not? Well, the obvious answer is "most people can swim with a lot less training than it takes to, say, learn Karate." Even more to the point though: if the game is (somehow) about professional swimmers (I once ran a game where a PC was on a high school volleyball team) then you want the skill to differentiate skill levels in competition. It's fine for that. If a PC wants to be a really strong expert swimmer then, okay, invest the points in the Sports Skill--that's what it's there for--but not only does a strong investment fail the reality check, it fails the good game design check. Forcing anyone who might go near water to spend 2-3 CP or risk death is just a bad call ... and, really, vanishingly few characters in fiction are distinguished "by the fact that they can swim."

So How DO You Purchase It?
Well, a look at the all-knowing Wikipedia under Swimming suggests that in the UK kids who can't swim 25 meters get about 10 hours of remedial training. That isn't much in the grand scheme of things (further pointing out that it shouldn't be a skill for most people). Let's go to the book and see if there's an example.

I cite the limitation "Lousy Driver" (pg 83, I looked it up). In game-worlds where there is an assumption of competence then the lack of the skill becomes a negative point Trait (a defect). So, say, for modern day it's -1 for Can't Swim. Something like that.

In game worlds where there is no presumption then it could be a skill (but probably not an Expensive one unless you are a sport swimmer). So say it's a Normal-cost Skill for medieval Europe and an exotic one. A Level 1 ability in it would give basic survival level swimming. Higher levels would make you a little faster, give some resistance to negatives, and so on.

Alternatively you could just make it a 1 CP Trait "Knows How To Swim" and call it a day.

What About The Sport Skill?
A look at a couple of competitive swimming pages show that there are some special skills around the start, the flip-turn. Without further research, I would assume that the skill there gives advantages measured in seconds or less to the better practitioner. There's also military swimming mentioned which (I assume) trains for holding breath, intel-gathering, and setting up explosives underwater. Maybe you wire them to the trained dolphin or something ... I don't know.

Anyway, all of this is handled by the Sport Skill. In a competition the characters will be having a Drama Roll session to see who can generate the most SPs. Having a roll requires the skill (if you just had the Normal Skill you probably couldn't get more than 1SP for making it--probably at a -3).

So that's still an option if you want to play Michael Phelps.

-Marco (I'll answer his other question later)

Battle Royal ... and Swimming ...

Testing Now
In the background the simulator is running tests for "Beak." I've done several versions of Teeth so it wasn't clear to me how to (if at all) make "Beak" any different. Firstly, part of my vision for JAGS is that more distinction is better. It's better to have three different "striking martial arts" than two or one. It's better to have five different tracks of Fantasy Warrior ('Fighter') than four--and so on. Now, this can only be done when  the difference is actually mechanical (if a Beak simply looks different than Jaws it doesn't count).

I settled on the idea that when a sharp, hooked, beak attack scores a vital hit (4+) its damage-bonus doubles before damage is resolved. I looked up "Beak" in Google and determined that raptors have hooked beaks for tearing flesh--as that's what I was going for, I decided that being extra-deadly on a vital hit would work well enough. So: testing.

Battle Royal
I did not run any tests this weekend but E did. Specifically, he was doing a "battle royal" of all-vs-all for the various martial arts in JAGS (all the ones that have been simulated anyway). These are: Karate, JiuJitsu (also could be Jujitsu or Aikido), Kung Fu, Tae Kwon Do, and Boxing. Here's how they break down:

  • Karate. The hardest hitting art over all. It has poor blocks compared to some other arts. The strategy involves kicking at range and then punching if the attacker gets in close. We have a Karate character who is cross-trained in "Ground Fighting" to see how he did.
  • JiuJitsu. A Grappling art. It doesn't hit hard but blocks well and gives a special Throw maneuver as well as better grappling. There are two strategies: THROW-RELEASE-THROW and THROW-GRAB-MOUNT-STRIKE. The first one simply has the character get in and throw the target whenever it's his turn. The target will probably get up and hit back (or strike from the ground). In the second case the JiuJitsu guy will close, throw, then apply a mount and strike until the target is pulverized.
  • Kung Fu. Kung Fu gives a good block, a little extra damage with strikes, and they are at -2 to be blocked. This goes a long way towards circumventing target's defenses.
  • Tae Kwon Do. TKD hits slightly less hard than Karate but gets better blocks against kicks, and one "cheaper" (in terms of action points) kick per round. It doesn't block well.
  • Boxing. Boxing gives a powerful punch, an even more powerful 1x-per-round Cross, no kicks (no Medium Range attack) and a good-but-not-great block. It also makes you tougher. You get a few more damage points and higher Hurt Condition, and so on.
Our current feeling right now is that (a) all the arts should be basically equal (since they are all the same points) and (b) it's okay if, for example, grappling has a slight edge over striking but that goes the other way when the striker has a small amount of cross training.

Here's more or less what happened ...

Jujitsu: The Mount Wasn't Good Enough
The Throw-throw-throw Jiujitsu guy was winning way more than the Ground-'n-Pound Mount attack. We did some detailed debug and found out that Mount, as per the rules, simply wasn't bad enough for the person being pounded to make up for the efficiency of the throw-throw-throw attack. We added a -4 Damage Modifier and -2 Initiative reduction for the character on their back and it brought Jiujitsu's Ground-n-Pound game back (NOTE: the real test will be the submission hold game but we're not simulating that yet).

Boxing Sucked
It turns out having no medium range attack sucks when your opponents do. It also turns out that hitting almost as hard as Karate but having no grapple game sucks against a grappler. The difference wasn't bad about 38% Percent Of Victory against Karate and TKD's 50% but taking a 12% bath so you can be as cool as Rocky isn't part of our vision. The toughening aspects of Boxing didn't do enough and the "roll with blow" damage reduction didn't help enough either. Here are some thoughts:
  • Tougher. Boxing (at Level 3) right now gives +1 Damage Point. It'll probably give +2 or even, maybe, +3.
  • Footwork. Boxing pros are really, really hard to hit. Maybe Boxing gives you a -1 to be hit by punches? Or even a -1 to be hit by all strikes?
  • Better Roll-With-Punch. Right now the special move doesn't work too well but what if it was improved?
Kung Fu Needs Work
When we boosted Kung Fu's block from +1 above skill to the (now stellar) +2 above skill it was pretty much dead-on against the strikers (whose blocks didn't work well and boy could it block them). However to beat a grappler with strikes you have to wipe him out before he lands the mount--you have limited time. Kung Fu's lack of hard-hitting damage hurt it there. So we are considering what to do.
  • Counter-Strike. Coming up next will be testing a Kung Fu where, when you block an attack (maybe when you block by 3 or more) you get an immediate strike regardless of your comparative Initiative (or, maybe you have to be within 5 Init of the attacker?). This will make Kung Fu battles more flowing and should give a minor but significant bonus over the striking arts. I'd like to see Kung Fu, if it wins 55% against Boxing or Karate lose 45% against Jujitsu or Wrestling. Maybe we can make it come out like that.
  • Grapple Defense? Kung Fu has some funky stances and has a lot of leg-power. Maybe they are harder to take down than other striking arts? Remember: realism here is secondary to balance so all kinds of options are on the table.
Tae Kwon Do Is Surprisingly Balanced!
It's kind of a "kicking version" of Karate and, as such, we were pleased that it came off even despite having a lot of small differences. 

Next I'll do Swimming. HUH!?

Friday, July 9, 2010

Problems Arise

This blog is a continuing conversation--you'll want to read from the bottom up for it to make sense.

Testing Today
I'm still (slowly) running through bio-weapons Teeth. As you'll see in the post what happened with the tests I ran last night we suffered a setback. A mildly interesting setback.

Juijitsu and Karate
The elation at how "balanced" Jiujitsu and Karate were vanished upon closer inspection. We'd had the karateka trying to "get up" every Round against the Jiujitsu guy and failing (his roll to get up was a 6-, something like 10% chance of success). This cost him an attack. When he just didn't even try and struck "from the ground" his victory percentage was something like 70%. Too high.

However, this was compounded by the fact that the grappler wasn't actually doing a "Ground and Pound" attack. A Ground-and-Pound (G'nP) attack is a strategy so-named in Mixed Martial Arts (MMA) where a grappler (often a wrestler) takes down a target and holds them on the ground and then, well, pounds them. The target, usually a striker, lacks the wrestling skill to get up and can't effectively strike from his back (as opposed to the wrestler--who may not be a good striker in general but can rain down hammer fists and elbows and such from a top-mount position without much difficulty).

We did some testing and determined that the key to making this battle actually work was that we had to make the simulator handle a "Mount" move. That's complex--but it's correct. The strategy for a Jiujitsu G'nP attack should go: Close->Throw->Mount->Pound-Pound-Pound. This could take a few Rounds and gives the Karate guy time to go Kick-Kick (at range)->Punch-Punch-Punch-Punch (etc.). Since the Karate guy does a good deal more damage the question becomes can the Karate guy seriously injure the grappler before he's pounded.

What Is A Mount?
Here is a visualization:

In MMA Combat a Mount Position looks like this:

The deal is this: you can "begin" grappling with a Grab (the 'least' grappling) or Grapple ('more' Grappling). You can then try to "add a position on top of it" (in this case a Mount). I'll note here that there are a few ways to get a Grab (you can Throw, which also knocks the person down, do a tackle or take-down which can result in a grapple, or do some special moves like a wrestling shoot which can result in a grab or grapple as well as other effects).

It's also clear that to perform a Mount the target has to somehow be taken down (this is accomplished in the simulator example by a Throw--which also establishes a Grapple).

What We Have To Get Right
Right now, in the rules "as written" (meaning as we interpret them from the book and as the simulator, where it is meant to, will simulate them) we believe that there is a good case that the situation is 'reasonably balanced.' Being taken down isn't good for the karateka and when we implement the Mount the Jiujitsu guy will almost certainly be able to get it (and since the karate guy may 'resist' that'll cost him extra action points (REA) degrading his striking or blocking even more.

What we have to get right is the amount that the Mount helps. What's "in the book" may not be enough (or it may be too much--but we don't think so). This will lead to a change in core rules.

The Other Thing That Didn't Work
The other thing that I ran into is this: the damage for Monster-Jaws was, by our method, way, way, way to high. What do I mean by that?

Monster Jaws: the attack is Short Range and hits for PEN damage. It's a "bite attack" meaning that it can "keep what it hits by" and Worry the target twice a round once the teeth are locked on. The 'Jaws' part means it attacks with the initial bite only once per Round and the 'Monster' part means it has, each round, an activation roll of a 9-.
This means that on a given Round (Round 1) the Monster-Jaws-Only character makes an activation roll. If the roll is a 9 or less (about 40% chance) then hey, he gets one shot to hit with his mighty jaws. If he hits (and hits by 4+) he keeps it and worries the target twice a round thereafter. If he misses--or doesn't hit by enough--then that's it for that Round (he gets an ineffectual punch too but it's ineffectual). If the character does not make his activation roll (60% of the time) he only gets two ineffectual punches and gets clobbered.

This happens again every Round.

Now, in the "wild" it won't be like this: nothing is likely to have nothing but Monster Jaws. It'll be Claws, a Toxic Stinger, and Monster Jaws or something like that--but remember our testing shows that we can give a target nothing but that attack and get a decent value for the attack. That's good.

It isn't perfect.

The problem with just running the simulator is that you run  it and it, well, runs. In this case our playbook of trying to get the Monster-Jaws-Only character to 50% wins against his peers came out to doing 90 PEN with the bite at 16 AP. How over the top is that? A sword at 16 AP does 11 PEN.

It's so over-the-top that any single hit--even a bad one--obliterates the target. The reason our methodology was pumping this out was that even with our "less aggressive normalized herds" strategy the battles were short enough that the bite might do infinite damage and it still wouldn't win more than 50% of the time because the chances of it getting a bite off in the time frame that our characters could survive were pretty much about 50-50.

I discovered this when I looked at the output and went "What The Fuck!?" and ran a detailed battle (several) with the debug turned up and saw that, yes, in the fights the Monster-Jaws-Only character lost he lost because he never bit anyone. In the ones he won? He bit the other guy once.

What We Need To Do With This
Our test characters are 50% offense, 50% defense. This is because we are (a) trying for a "pretty" distribution of points. Having a character who is at least theoretically 50/50 makes some of the math easier. The other reason (b) is that we're sort of looking at a "worst case" scenario where a PC has built a very aggressive character (we think some PCs might well be more than 50% offense, including some reasonable characters like 'Vietnam Infantry Guy' who has an assault rifle and no armor).

With these builds we see the fights take about 2.5 to 3.5 rounds at the upper end and in order to understand what Monster-Jaws-Only might really be worth we probably want to see a fight that takes, on average 4.5 Rounds.

What I did was give that character a 1/4th AP Super Punch to use when the Jaws weren't working and I got some numbers that "look" reasonable. If I have the time and energy I'll go back and look it over again with a better distribution of APs to give our test characters better defenses (and therefore more staying power).


Thursday, July 8, 2010

Caught Up

This blog is a continuing effort and many posts build upon the previous ones. I advise you to read from the first post (at the bottom) up.

What I Did Today
Having found my DQI (Data Quality Issue--a mistake) from yesterday I spent today running the simulations for Jaws (big teeth) and Saw-Toothed (a kind of bite attack where the "worry" action does more damage than the bite). It was slow going but since the simulator runs int he background it's just a matter of setting it up and letting it go. By the time I should be finished I'll be caught up to where I wanted to be 2 days ago.

Something We're Hoping For
Many of the attacks will share the same "configurations." An example of a configuration would be "Fires once a round on every other round starting with the first one (this is called POWERHOUSE)." Another example would be "The character must waste a 5 REA Medium Action doing nothing and may then use this attack as a standard attack (this is called ACTIVATE)."

It is our hope that we will discover that different attack types respond the same way to a given configuration. That is, let's say configuration POWER HOUSE multiples the effective damage of a Power Blast (Ranged Impact Damage) by 3.75 (so a Power Blast that did 10 IMP damage with a standard fire attack would, in the Power House configuration do 38 IMP damage). If this is true--and it's true for "gun damage" (Ranged PEN) and punch damage (Hand-to-hand IMP) and so on then we don't have to test it for weird things like Disintegration and so on--we just give them the modifier.

The deal is that we'll need to do a lot of testing before I'll take it on faith that a specific modifier is universal. Disintegration (which, right now, is something like "Ignores all armor and, if the damage forces a CON roll, it makes it at -5") is very, very different than Impact Damage. It's much much worse and probably has a sort of steep "cliff" wherein a small amount is not especially dangerous (as it does not force a CON roll against opponents with anything invested in Damage Points) and then suddenly becomes hugely dangerous once it does force a CON roll (maybe it'll force the CON roll at -3 for animate opponents or something. I'm not sure yet).

These exotic attacks will probably need more testing anyway to ensure that they aren't improperly designed in the first place (negatives to CON may prove bad for the game).

What Else We Learned Today
In a second set of testing, E. has the simulator working for Jiu-Jitsu grappling--at least part way. We now allow a Jiujitsu fighter to block-throw and then hold the opponent down and pummel them. In the future we may allow him to go for a joint submission.

We tested this against a karate-fighter and, in theory, the battle should be 50-50 as the number of points these skills cost are the same (NOTE: I am very, very familiar with the history that Jiujitsu has against Karate in real life. That's another discussion). As we want these to balance at the same cost levels we were pleased to discover that the match went 55 (Jiujitsu) to 45 (Karate). It's not perfect but for something that was not developed with the simulator it's well within tolerances.

Even better: while the Karate guy could not get up and did not have ground-fighting skills his higher damage-per-strike meant he got his licks in anyway: it wasn't just a blow-out. I couldn't hope for anything better from the system.

How Come We're Not Letting Grapplers Beat Strikers?
One of the great realizations of modern mixed martial arts tournaments was that "pure grappling" was superior to "pure striking" inside of ring conditions. It is also believed, by many (including me), that people who train grappling often train at a higher intensity level than those who train striking (most karate classes are not training for a full-contact ring fight where as most grappling arts are). I think this is important: it's not easy to knock someone out with one shot and once a pure striker is down they lose a lot of options against a skilled grappler.

So what? Why don't we reflect this?

Well, (1) we kind of do. The "Ground and Pound" attack works pretty well in JAGS if you can pull it off. That's what we're seeing in the JiuJitsu-vs-Karate fight. If the Karate guy had a little more cross training (i.e. had something that gave him "Ground Fighting" skill) he'd do a lot better. This "reflects reality" pretty well as I see it. Secondly (2) We're less interested in conforming to a specific view of reality (even if it's one we generally hold) than in making a fun game. In many, many fictions highly skilled (expert-level) karate characters are quite dangerous to just about anyone and we want to reflect that. The way the battle goes (with the karate fighter getting taken down and then fighting at a disadvantage) is pretty much exactly what reality says happens (the difference here, so far, is that the grappler is more or less standing and opts, for now, entirely for strikes rather than submissions).

What's out of whack is how much damage the Karate guy can do in the exchange: it's probably a little higher than real life dictates. Also out of touch is the time-span: we know that in MMA rings although some fights are over in 12 seconds most last more than 60 (10 JAGS Rounds). That's way, way more than most JAGS fights. We also know why: there isn't the constant level of engagement that a JAGS fight has--there's a lot more circling.

We actually have a sketch of rules that approximate that: if you don't win Initiative by enough more than your opponent (and its secret) then when you come in to strike you are pretty darn open for a powerful hit. This rule would tend to have both parties "circling," waiting for a "really good initiative roll" or otherwise trying to draw an opponent in (or just going for it and taking the risk). Our examination of this has led us to believe it wouldn't lead to better games so it's not in there.

For similar reasons we don't subscribe to grappling's observed dominance in sport-fighting.


Wednesday, July 7, 2010

It Doesn't Always Work

Read the blog from the bottom up or it won't make sense.

I didn't have a lot of time to work on JAGS today but what I did do hasn't panned out so well. I'm trying a variant of "Jaws" (teeth) that does more damage once clamped down than the bite-and-release does. I'm calling it Saw-Toothed.

The problem is that the attack-mode has two damages and trying to get the numbers right and have them stay constant or semi-constant throughout the progression (16 AP, 32 AP, 64 AP) is causing me problems. I'll add to that the fact that the attacker's decision to bite-and-hold is something I have to set: in JAGS the amount that an attack "hit by" is important and for a PEN attack a hit by 4 or more is considered a hit to a 'vital' location doing more damage than a hit by 0-3.

Currently I have the simulator set to grab-and-worry on any hit by 4 or better (this stops the attacker from "keeping" a very bad hit). However, in the case of Saw-Toothed it might pay off for the attacker to keep any hit they get (after the grab the target can't block and the strikes don't miss). I haven't tested that yet. It's time consuming to set up the test-run and I'm having enough trouble getting a ratio between bite and bite-and-hold damage to hold steady.

So far it looks like 1:4 is maybe reasonable. That is, if the initial bite does 12 then the worry attack does 48. This is pretty darn high, however so I need to see if the results I'm getting are good. Most recently an 18 PEN 1x per Round Bite with a 72 PEN worry (2x per Round) won 45% against the 16 AP Herd. For comparison, a 1x Attack Bite that scored 50% did 24 PEN (Bite & Worry are the same damage). This means that a 6 PEN deficit would, by the simulator, give us a  48 PEN increase. That seems awfully high to me.

The letters DQI stand for "Data Quality Issue" and it's a generic term for concerns about any result that seems out of whack. When I get a weird result (i.e. I reduce the damage and the % of wins goes up--or I increase the damage and the % of wins doesn't change--or whatever) then I have to immediately look for DQI's.

This is things such as "did I update the Herd Spreadsheet correctly?" It can also be things like "Did I render this attack properly in the Attacks Tab?" On the other hand, the worst case scenario could be "We have a bug in the simulator." It's happened before. A bug now could invalidate all the testing to date if it were pervasive and bad enough. The good news is: all of this has been looked at at least a little. The bad news is: none of the minor bits of the simulator (such as the bite-and-worry code) have been looked at hard enough. A single run (one battle against another) creates hundreds of lines of output when DEBUG is set to maximum level. Our trial runs are groups of 5000 combats so if something weird only happens once in a while we have to determine where by reading the code--not by watching it happen--and then put in some breakpoints or extra debug or something.

This means that if what's going on with Teeth happens to be a DQI? It means I could be debugging this for a long time.

Post Script
I found a DQI in the list of attacks that would explain some of the problems. I'm testing it now.


Tuesday, July 6, 2010

Philosophical and Technical Questions

Read this blog from the first post up or it won't make sense.

I'll save the discussion of what kinds of modifiers for ranged attacks we're testing to talk about something else today: Philosophical Questions vs. Technical Questions.

Testing Right Now
We'll get this out of the way. The simulator is cranking away on various kinds of Teeth. Teeth can do a complex bite, bite/grab, and bite-grab-worry chain of attacks. They tend to be close ranged weapons and attacking with teeth leaves you vulnerable (it counts as a "cross" strike making you at +1 to be hit and -1 to block until your next Round). There are at least three different sizes of Teeth and I'm testing away.

Philosophical vs. Technical Questions
We're using a computer simulator to test things but that doesn't tell us what to test against. When trying to figure out what precise test to run a lot of the effort goes into creating a sort of "controlled experiment" (you can read the previous notes to see how we've evolved our testing strategies. Even the strategy, however, won't tell you everything.

Right now we have a decent concept of a "Normalized Herd." Remember that's a group of 20 characters who are highly simplified and represent a range of attacks and defenses we think are "representative" of 'most games.' You get attacked by a sword guy, a gun, guy, a super-punch guy, and an energy blast guy. They have mixed armor and DP defenses. It's simple--but it's indicative.

So now we run into the hard questions that sort of come down to "what is the 'average' RPG combat like?' If you've played several different games for several years with different groups you can see how this would be a hard question. Here are some common types we've identified:
  • The Boss Battle. This is a "hard" combat of the group against one really tough opponent. The Boss has to be able to take a lot of incoming fire (the whole party) and dish out enough to hurt the characters but probably not obliterate them. We think these are an important part of Fantasy and Supers combats (although they can certainly exist elsewhere)
  • The Peer Battle. This is a one-on-one battle where you are facing an equal. We think that in a lot of games these are fairly rare in a lot of games but are common and expected in Supers combat. NOTE: there are some game types such as the modern-day 'cops' style game where we think this would be pretty much 'almost every combat' (since no one has many Archetype Points--although the PCs may be more combat capable than most). Those don't count for as much: we're looking for AP-based characters where we'd see "wizard duels" and the like.
  • Goon Battles. A goon is described as a lower point character who hits about as hard as a PC. A goon is a real threat although they seriously lack the staying power. An example might be a character who has spent almost ALL his points in offense so as to hit as hard as higher-point characters. D&D 4e's Minions are examples of this. We think this is pretty common in most games. It might be "the most common."
  • Mook Battles. A mook is simply a lower point character--usually so much lower that they, even a group of them, are not much of a threat. We think this is going to be common in a lot of games but not so important as to balance since the mooks are expected to lose with only minimal damage to the PCs.
  • Skirmish. For want of a better name a Skirmish is a battle where the PCs are facing a smaller number of opponents that are pound-for-pound tougher. A party of 5 PCs might face 3 Dire Cows or something. This is a classic fantasy-monster configuration.
If you look at these battles you can see that what we're thinking about is "where should a monster or NPC spend its points to try to create a certain type of experience." It also asks "what does "balanced" mean for a given character? If you are one of 5 PCs fighting a boss what is expected of you? What about if you are fighting goons: is it the same thing?

This leads us to the crux of both kinds of thinking. For the Technical Questions we can ask things like what happens when we fight a 2-on-1 battle against a character with low defenses? When we see that Damage Points are way, way less effective against 2-on-1 opponents compared to armor we can start making decisions about how to address that (solution: allow characters with no armor or Force Fields to buy "extra Damage Points" cheaply).

These are fairly easy questions. Although we have to figure out how to set up the simulator for that (in our case, I had to modify the code to handle multiple opponents) once that's done we can determine what kinds of rules lead to outcomes we like.

Philosophical, questions, on the other hand, are much harder.

Philosophical Questions
Philosophical questions are things like "how many attacks are ranged vs. hand-to-hand" or "how many battles are 1-on-1 vs. 1-against-many?" These are difficult questions as there is no right answer--not for a universal game system. Most of this is up to the group and the GM. Some of it will be up to the genre and the specific PCs (Luke Skywalker is more likely to get in a melee battle than Han Solo). 

This is important when dealing with things like:
  • Area of Effect Attacks. If you are fighting 1-on-1 at range an explosive attack can be good in that it doesn't miss but it has a whole new appeal when it can take out 10 guys who were going to fire at you next round.
  • Short Range Damage Fields. If you have a power like Electrified Body that hits people trying to hit you then it's really good against bare-handed attacks and won't score so well against guns or swords. If your simulator (as ours does) includes only 5 HTH combatants then the score for how good that defense is will be lower than if the simulator had half of the opponents come in bare-handed. On the other hand, if you are playing a martial arts game where all combat is HTH then the cost will be way too low (we doubled the tested value).
  • Attacks That Penetrate Armor. We assume that 75% of our combatants will be some-how armored. If an attack is good against armor (but doesn't hit much harder if the target is not armored--such as an Armor Piercing Attack) then in our simulator it's going to look pretty good: for a lot of those characters their primary defense (armor) is going to get degraded. However if you are in a game where many people are NOT armored (say a Chi Martial Arts game) then that extra bonus won't be nearly as good.
There simply is no good way to test all of these conditions--they're contradictory (if we assume most battles are 1-on-1 then attacks that can hit multiple opponents will be cheaper than they 'should' be in games where most battles are many-on-1). What we can do, however, is make some assumptions, publish them, and think about what the ramifications are. Here are some of them.

The 'Supers' Game Is Real Important
A Super-hero battle is probably the most important one for "full spectrum balance" that we can easily think of. Take for example the host of modern-day style games (horror investigators, cops/robbers, and military guys). If the PCs are not extremely high points the combat aspects of these will be less important than other ones. On the other hand, if you are playing a super hero you expect to have a cool interesting fight against a peer and not just get cheap-shotted or one-punched out of the scenario (exceptions, of course, do exist).

Armor Is Really Good
We know that armor is pretty much the only JAGS defense that doesn't degrade (outside of negative Damage Modifiers and Negatives to be hit). Damage Points, Ablative Damage Points, Force Fields, and Power Fields all degrade over attacks so armor is the most reliable form of stopping damage the game provides. It also tests the most expensive point-for-point. We need to keep this in mind when creating traits: PCs should have access to armor but if we don't want it to be overly dominant then we need to keep an eye on how it is being sold.


Friday, July 2, 2010


Read this blog from the first post (at the "bottom" of the page or pages) up or it won't make sense.

Right now I'm (painstakingly) testing a massive suite of attacks using the most recent realizations about costing to try to balance them (I think the discussion around 'balance' is for another post). While that test suite is being run in the background (and it'll run all day long) I want to talk about why it's so big and what goes into it.

This post will stick to bio-weapons. These are your basic claws, teeth, pincers, and so on. Most of these do PEN damage and all of them add to your basic Strength/Build scores so that larger things hit harder. This is all pretty straight forward but let's take a look at these basic things ... and how they get modified.

Kinds of Bio-Weapons
Let's start with some of the common categories:

  • Blade/Talon. The most basic bio-weapon is the "blade/talon." This is essentially a broad-sword (which may or may not be a literal sword or might be some kind of unusually large natural blade or tusk or something). It's notable in that it does PEN damage. usually has some kind of "reach" (a punch is Short Reach, a sword is Medium Reach). Straightforward.
  • Teeth and Jaws. This kind of attack also includes the "pincer." In this case the attacks can do something called a 'worry.' A Worry is the term for what happens when a dog (or other animal) bites something and then shakes it, the blades sawing through the flesh. In game terms when the attack hits the user can choose to "hold." If the hold (a grapple roll) succeeds then for 1 action thereafter they can attack again with no roll to hit, no chance to be blocked, and the same hit-by modifier of whatever the original attack hit by (so a great hit "to the throat" is more likely to result in the attacker choosing to try to keep it than a poor hit "to an arm"). I really love the simulator because this sort of attack is devilishly hard to cost properly just using math (there are all kinds of decisions that have to be made and the attack is quite complex along several parameters).
  • Tails. Tails are usually expensive (in terms of action-points) to use (costing 6 REA instead of 5) and are (typically) hard to block (-1 to block/dodge). 
  • Toxin-Delivery. This is another case where we were boggled trying to cost it by hand. A toxin-delivery system (fangs, stinger tail, poison dart or quill) is complex. Firstly there is the toxin itself which may do straight up damage, some manner of incapacitation, or both. Toxins have differing 'strengths' (called Intensity) and will work better against a character with tons of armor but few damage points than against a character with lots of damage points (but maybe no armor). On the other hand, they have to get the toxin into the target so there's the factor of "injection" (this problem does not apply to 'poison gas clouds'--but those are treated separately). In this case the attack must "penetrate the armor" doing at least 1 point of damage. This is complex to figure out: how likely is a toxin attack to penetrate? Well, that determines on who is throwing it (if they are super strong, the chances go way up). Trying to figure out who was "likely" to have a toxin attack nearly broke us--figuring out how to factor strength of toxin against a deadly carrier attack (a stinger tail that hits hard enough to potentially kill in its own right) was possibly beyond us. The simulator helps a lot here.
  • Defensive Spines. This can also include things like electrified or burning body. In this case the trait is "passive" in that the character doesn't usually use it--instead it "goes off" when someone hits you (think of a porcupine). Again, this is complex: the character is usually assumed to have some back-up attack and the damage isn't usually dealt "at range" so this kind of defense is meaningless against a gun-using opponent. How we'd factor an attack that fed off the other person's action points was a tricky question. The simulator gives us a value but whether it's correct or not is a matter of philosophy (the simulator splits ranged and HTH attacks evenly. If you are playing in an all-hand-to-hand game like a martial arts tournament for mutants then the value will be much higher than the simulator tests).
Now that we've looked at some of the bio-attack types, let's talk about the common modifiers. A modifier is a change in how exactly the attack is "thrown" or what conditions go along with its use. Usually these have some specific intent when it comes to character design.
  • Monster. We're using this term as a place-holder. It means the attack will have an "activation roll" each round of a 9- roll (just under 50%). The purpose of this modifier is to allow the creation of monsters who have several attack types but can't always use all of them each Round. Imagine something like a giant scorpion with bladed pincers and a very dangerous poisoned tail. The monster creator can give the tail an activation roll so that most Rounds it'll use its blades but sometimes (and you can never be sure when) it has the option of stinging with the tail. This mixes up the combat flow and, we think, can create excitement.
  • Charge Up. A charge-up attack (as noted before) has a 1 or 2 Round period before it can be used. This means that the character cannot "open" with the attack (and can't use it every round thereafter, it re-sets after each Round of use). Depending on the configuration it may or may not be able to be used more than once when it is active. This allows the character to have a very powerful attack that can't be immediately deployed against an opponent: the character may need to fight defensively or attackers may realize they have to work quickly to try to avoid it.
  • Easy To Dodge. We haven't done a lot with this to-date but it's under consideration. In the Fantasy games 'rogues' may have a high dodge defense as opposed to everyone else's block defense. A good Dodge is purchased (usually) with the somewhat exotic Acrobatics skill and because it costs like a combat skill you won't see too many 'front-line fighters' with it. On the other hand, it'll appeal to characters who, due to some specific Archetype Traits get some "free points" in Acrobatics or otherwise want to invest there. If we give some modes of bio-attack a deficiency against Dodge defenses ("slow and clumsy") then we allow the (usually more lightly armored) acrobatic characters to survive them better. This creates a (we think beneficial) situation where a monster with a powerful attack ready to go will have to choose between aiming it at a heavily armored front-line fighter or going against an more lightly armored (and possibly less tough) agile character who will not only be harder to hit in the first place but gets extra defenses if they take a Dodge defense. This should tend to aim more powerful attacks at the tougher 'tank' characters. We think that's good.
For ranged attacks we have several other configurations we are playing with and those deserve another post. We are also thinking about how these modifiers might apply to "martial arts Chi fighters" or HTH-Bio-Mod Combat Cyborgs. Would there be any real differences? We're not sure but if there are some good ones we want to consider the ramifications now rather than later.


Thursday, July 1, 2010

Testing and Re-Testing

Read this blog from the first post up or it won't make sense.

I've reached a 75% mark on Power Blast and have come to the conclusion that ALL the tested HTH attacks and General Archetype Traits (which I have already "completed" before this blog started) need to be redone. This is a recurring pattern with JAGS Revised Archetypes and it sucks ass. Here's what's to be done--and why.

The analysis of Power Blast showed us that all our costs were essentially some kind of wrong. The reason was that there is no real "cost per point of damage" for Power Blast and, probably, not for other attacks as well: there is a "start up cost" (Level 1) and then an on-going cost which may or may not be balanced out to a hypothetical infinite number of points (we have a clever way of not needing to deal with an infinite number of APs but I'll cover that later--the basic mechanics should still work, however).

So what I'm going to do is take a list of attacks such as:

  • "Strength" (HTH Impact) and its "child" attacks like Thunder Tail (ranged STR with a negative to block and a higher REA cost per swing), or Ram's Horns (blunt attack used with a charge maneuver) and re-calculate them using the 16, 32, 64 methodology.
  • "Sword" (HTH PEN) and its child things like claws (Close Range HTH), Jaws (a "Worry" attack where a single hit can be used to shake the target dealing more damage without having to re-roll to hit, and so on.
  • "Gun" (Ranged PEN) and its child attacks such as Ghost Bullets (ranged PEN that ignores armor), and cutting beam (a ranged PEN attack that ignores armor on a successful PEN roll) and so on.
This is a lot of testing. Each test for a given attack type (much less the numerous child attack types) could take a couple of days. This could be months of testing and re-testing.

Why would I subject myself to this? Well, the basic reason is this: the game I want to play is one that will (a) give a great deal of satisfying latitude in character generation (i.e. there are many options and they are relevant to the mechanics and the experience of playing) and (b) I want the "point buy system" to create 'balanced characters' without me as a player (or as a GM) doing a lot of policing to keep it balanced. If someone takes armor piercing claws and superior reflexes and precision striking? I want the native costs in the book to all work out so that character "punches at his or her point level."

The simulator as a tool, a clear set of goals, and a scientific approach towards analyzing the data seem to be able to meet that goal. In other words: I want this project to be the best it can be and, although it's a depressing and often overwhelming amount of work, it seems to be closer and closer within reach (I am so happy I didn't say "fuck it" and publish the book 2 years ago--we would have created numerous other JAGS source books based on it and the present catalog would be further and further out of date and the rules would be more and more broken).