Request for idealized guinea pig for mathematical modeling.

[Varith]

Greetings.

I know I am new to these forums, but I have been playing and using hero system for the better part of two decades. I have spent time reviewing various threads on changes to the system, balance issues, and mathematical analysis of things like the killing attack vs the normal attack.

I love the idea behind what I am seeing, but I think the true evaluation of an attack or defense power is how it affects final victory in combat. To that end, I have written a program to run automatic simple combat between characters and record wins/losses as I vary the stats, attacks, etc. So, in about 30 seconds, i can run an experiment with 100,000 battles or so, and tabulate the results. There have been a few eye openers.

But the question is, what is a 'standard character/guinea pig' to plug a power or defense into?

My own starting point, and one I suspect that the original designers used decades ago, looks something like this: (no mental stats yet, just working on non-ego powers and such)

Str: 23
Dex: 23
Spd: 5
Con: 23
Body: 15
Stun: 39
End: 46
Recovery: 10

Additionally, my balancing point is 60 point attacks (12 d6 EB, 4d6 RKA, RKAs with rule changes) countered by 60 character points spread equally between PD and ED. (amount of resistant vs non-resistant defense varies per experiment)

If anyone has thoughts as to the idealized guinea pig framework to evaluate combat powers, I'd love to hear your thoughts and rationalization behind them. When I get a bit more time, I will post some of my results and conclusions so far.

Varith

[Varith]

The reason that I posted this in the 6th edition forum is that there are qeustions about changing the basic mechanics, such as the killing attack resolution. Netzilla did a very good job of looking at the breakdown of damage class and defense vs. expected damage. I am hoping to shed more light on this and other questions.

Let me post some preliminary results:

I started by taking my guinea pig template: Str-23 Dex-23 Spd-5 Con-23, Stun-15, End-46, Normal PD-15, Resistant PD-10, Total PD-25, Normal ED-15, Resistant ED-10, Total PD-25. Then, I gave the first one a 12 damage class energy blast, and the second one a 12 damage class RKA, using classic damage resolution. Given all of the screaming about the stun lotto, I expected that the killing attack would, statistically, be victorious.

I ran 100,000 fights, and ended up with the following results in about 40 seconds:

Energy blast dude: victorious 52.5% of the time
Killing attack dude: victorious 47.5% of the time.

Victory being defined as not losing, when losing means going negative one's starting body, or going below zero stun. The very primitive AI will choose to dodge rather than attack if there is not enough endurance to fire a shot without causing self inflicted stun. Recoveries are taken post-12, and at this point, bleeding while negative body is not yet implemented.

So, although there may be at times problems with the stun lotto, if I am going up against characters with roughly comparable points, I would actually have a slight preference for the energy blast, assuming that my template is close to the average type of characters flying around.

I may not have the proper 'central guinea pig' around which all of champions resolves. I do believe that this style of actual combat simulation may solve a number of questions as to whether any key mechanics need changing.

Sincerely,

Varith.

[Tonio]

Just checking... but did you include Stunning in there? As I understand it, one of the problems with RKA vs EB is that RKA will Stun more often, which would have a serious effect on combat (no active defenses, can't attack, lower DCV, etc.). Also, remember you're prevented from recovering from Stun if you take damage while recovering, and that you can't Abort a Phase to recover, and you lose any Held Action (so you can't use that to recover, either).

[Varith]

Stunning is included in the program, so if Dude 1 scores a stun, Dude 2 loses his next action to become unstunned. The only recoveries that are taking place are the post-12 recoveries, and as far as I know, nothing short of unconsciousness or death prevents post-12 recoveries, (except the optional rules for hit locations, where an impaired Vitals hit stops post 12 recoveries). The primitive AI never tries to take a recovery as an action, as that would just open one up too much. Post-12 recoveries do not create any vulnerabilities, at least as far as I remember in the rules.

I will include stats on average number of stuns scored and knock-outs vs deaths as I continue to evolve the program.

[Tonio]

Stunning is included in the program, so if Dude 1 scores a stun, Dude 2 loses his next action to become unstunned. The only recoveries that are taking place are the post-12 recoveries, and as far as I know, nothing short of unconsciousness or death prevents post-12 recoveries, (except the optional rules for hit locations, where an impaired Vitals hit stops post 12 recoveries). The primitive AI never tries to take a recovery as an action, as that would just open one up too much. Post-12 recoveries do not create any vulnerabilities, at least as far as I remember in the rules.

I will include stats on average number of stuns scored and knock-outs vs deaths as I continue to evolve the program.

Oh, by "recovering" there, I meant recovering from the Stunned condition. If I get Stunned this Phase, and I already acted, I use my whole next Phase to try to recover from being Stunned. If, during that Phase, I get hit for damage (either BODY or STUN), I fail at recovering from being Stunned and stay Stunned.

On the other hand, if I haven't acted this Phase yet, I can use this Phase to recover (which probably means I won't be interrupted, since whoever Stunned me to begin with already used his action to attack me).

[Varith]

Because of it currently being a duel between two equal speed combatants, the only penalty to being stunned is loss of the next action. I do admit that in a major chaotic melee with multiple opponents, someone might take an opportunistic pot shot at a stunned character while he is trying to recover. In my model case, the two characters are always alternating shots (think Dex 24, spd 5 as dude 1, and Dex 23, speed 5 as dude 2, dude 1 and dude 2 trading stats every other fight). Because of this, there will never be 2 actions in a row by either combatant to both score a stunning hit and get a follow through attack while the opponent is still stunned.

I can't model everything yet. There is only so much one can do with 3 hours of visual basic coding time to put something together. (My program runs as an Excel visual basic for applications program, so it would be pretty easy to send around sometime and let people see it and mess with it.)

Varith

[Varith]

I had to revise the code slightly to include the rule of 1 body done = 1 stun done for killing attacks, which has a minor effect upon the results. As a result, the basic guinea pig with a 60 point attack, and 60 character points of defense (30 character points of resistant defenses), has a killing attack was almost exactly equal to the normal attack in combat effectiveness. Here is the guinea pig and results

Str-23 Dex-23 Con-23 Body-15 Spd-5 Stun-40 End-46 Recovery-10
Normal PD-15 Resistant PD-10 Total PD-25
Normal ED-15 Resistant ED-10 Total ED-25

Combatant 1 has a 12d6 EB
Combatant 2 has a 4d6 RKA

The fight was run 100000 times

Energy blast dude won 49.939% of the time. He died 617 times. (0.617%)
Killing attack dude won 50.061% of the time. He died 0 times.
Energy blast dude was stunned 0.60107 times per fight (that is about 3 fights in 5 he was stunned)
Killing attack dude was stunned 0.27201 times per fight (that is about 1 fight in 4 he was stunned.

Next, I cut the defenses in half, to simulate the games with high powered attacks, and low defenses (howitzers mounted on eggshells!)

New guinea pig, howitzer style
Str-23 Dex-23 Con-23 Body-15 Spd-5 Stun-40 End-46 Recovery-10
Normal PD-8 Resistant PD-5 Total PD-13
Normal ED-8 Resistant ED-5 Total ED-13

Combatant 1 has a 12d6 EB
Combatant 2 has a 4d6 RKA

The fight was run 100000 times

Dude1 down 36811 times. Dude1 dead 1792 times. Dude1 stunned 44137
Dude2 down 63189 times. Dude2 dead 0 times. Dude2 stunned 113492

Energy blast dude won 63.189% of the time. He died 1792 times. (1.792%)
Killing attack dude won 36.811% of the time. He died 0 times.
Energy blast dude was stunned 0.44137 times per fight (that is about 4 fights in 9 he was stunned)
Killing attack dude was stunned 1.13492 times per fight (more than 1 time per fight).

OKAY!!! notice this! In the case of lightly defended (but balanced defenses), the energy blast defeats the killing attack by nearly 2 to 1!!!! Also, the number of STUNS scored by the energy blast is more than 2.5 times as many as the killing attack!!!!

Last experiment for the night: Armored tanks with BB guns

I will take the base guinea pig and increase defenses by 50%. So....

Str-23 Dex-23 Con-23 Body-15 Spd-5 Stun-40 End-46 Recovery-10
Normal PD-23 Resistant PD-15 Total PD-38
Normal ED-23 Resistant ED-15 Total ED-38

Combatant 1 has a 12d6 EB
Combatant 2 has a 4d6 RKA

The fight was run 100000 times (this took ALOT longer to run)

Energy blast dude won 3.586% of the time. He died 3.501 times. (3.501%)
Killing attack dude won 96.414% of the time. He died 0 times.
Energy blast dude was stunned 1.27459 times per fight (more than once per fight)
Killing attack dude was stunned .00203 times per fight (laughable).


Aha! this is what people scream about!

My question to you, is this a sign of energy blast being inherently bad, or is this a sign of defense heavy campaigns?

More analysis to come,

Varith

[Hyper-Man]

Any plans to release a user friendly version of that program?

[Varith]

I hope over time to make it more user friendly... such as having all the stat values in an excel spreadsheet and check boxes for various options.

To continue...

I was startled by the high defense result. In an ideal world, wouldn't a balance between points spend in normal defense and points spent in resistant defense result in equal protection from the two attack forms?

I thought, maybe it's just a small variation in the balance between them. in the case of the tanks (60 point attacks, 90 points spread out in defenses) Let's try out all resistant defense and see what happens:

Guinea Pig: Str-23 Dex-23 Con-23 Bod-15 Rec-10 Stun-40 End-46
Resistant PD/ED - 30. Normal PD/ED-0 Total PD/ED = 30

Dude 1 has a 12d6 energy blast
Dude 2 has a 4d6 RKA

RESULTS:
Energy blast dude won 40.912% of the time! Stunned 0.8193 times/fight
Killing attack dude won 59..088% of the time! Stunned 0.07228 times/fight.

THIS I believe is the heart of the hatred of what has been called the 'stun lotto'. In my opinion, the 'stun lotto' is a symptom of a much deeper problem.

Postulated Ideal:
1. In an ideal world, spending points on normal defenses SHOULD defend against normal attacks, and spending points on resistant defenses SHOULD defend against killing attacks.

2. With balanced defenses against normal and killing attacks, the two types of attacks SHOULD be EQUAL in their ability to achieve victory.

3. This balance of the normal and killing attacks SHOULD remain relatively constant even as the attack to defense point ratio is changed.

if ANYONE disagrees with these SHOULD statements, I would LOVE to hear your reasoning behind them, and how the postulates SHOULD read for a healthy, robust role playing game that we will all love to play.

Observed Results (from experient).
1. In the case of balanced points in attack and defense (60 attack, 60 defense), and balanced points between the defenses (30 character points in PD/ED, 30 character points in rPD/rED (Armor)) the normal attack IS balanced against the Killing attack from the perspective of Victory in Combat (50%/50%)

2. In the case of attack heavy games (60 pts attack, 30 pts defenses), the NORMAL attack becomes far more effective than the Killing attack.

3. In the case of defense heavy games (60 pts attack, 90 pts defenses), the KILLING attack becomes far more effective than the normal attack.

I assert that this is a problem, and ideally should be addressed and corrected.

For those who do NOT think it is a problem, I would love to hear why, and why such a skew of normal/killing balance vs attack/defense points is a good thing for the game.

Sincerely,

Varith

[Steve Long]

To the extent there's anything here that's useful to the 6E discussion (at the very least it seems interesting from a theoretical perspective), it belongs in one of the standard threads, such as General Issues or Combat.