The great debate, this time with Java!

[Kdansky]

Re: The great debate, this time with Java!

 

Ok, take these with care. I added stun calculation for normal damage with hit locations, but I'm highly unsure whether there are bugs in it, as it is quite untested and I am not sober enough to not make stupid mistakes (well, if I was sober, I can still make them, I just make less).

NORMAL ATTACK . Stun Rule: normal damage
Plinks : 1.48 % plink ratio
Stun / atk: 11.51
Con< / atk: 0.23 %
Body / atk: 0.0

KILLING ATTACK . Stun Rule: 1d6 -1
Plinks : 44.29 % plink ratio
Stun / atk: 11.57
Con< / atk: 17.43 %
Body / atk: 1.45

KILLING ATTACK . Stun Rule: fixed x2
Plinks : 50.0 % plink ratio
Stun / atk: 2.91
Con< / atk: 0.0 %
Body / atk: 1.45

KILLING ATTACK . Stun Rule: fixed x3
Plinks : 9.25 % plink ratio
Stun / atk: 11.89
Con< / atk: 4.62 %
Body / atk: 1.45

KILLING ATTACK . Stun Rule: hit locations
Plinks : 26.44 % plink ratio
Stun / atk: 11.46
Con< / atk: 12.57 %
Body / atk: 1.45

NORMAL ATTACK . Stun Rule: hit locations
Plinks : 1.48 % plink ratio
Stun / atk: 11.53
Con< / atk: 0.01 %
Body / atk: 0.0

 

Lo and behold, hit locations are a disaster for Normal damage. You get a bit more average damage (like 2%), but you lose any chance of ever stunning, since you have to be extremly lucky to first roll exceptionally high on your 9d6 AND then not roll any of the x0.5 stun rolls on the Hit Location chart, some of which are very common rolls (6,7,8,15,16,17,18) and the x1 rolls don't help much either (9, 10, 11, 12, 14), whereas only the unusual rolls give you x2 stun (3,4,5, 13), except of 13, but that one is easily destroyed by the 6,7,8 combination with x0.5 stun.

 

So if you use Hit locations, you slightly nerf KAs and you nerf EBs at least as much. Nope, that does not help. Well, on second thought: Since EB will already never stun (0.23% is nothing), then reducing that to 0.01% is not much of a difference, even if it is a factor of 20. It's only 0.2% really. So yeah, if you use HL, your EBs lose a bit of punch, but the KAs lose more. Still not a sound solution.

 

Files: http://rapidshare.com/files/91831752/NvsK_kdansky_v4.jar.html

Time for bed, it's 2:35 here and I have to get up early. Hurray.

[Paragon]

Re: The great debate, this time with Java!

 

I have to admit, I haven't seen the KA problem as much as I expect to. Part of that, of course, is that I run more than I play, and I use a KA variant. When I play, I occassionally gaurd against it, too, pulling tricks with DR or whatever to minimize the impact of the STN LOTTO on my character.

 

Statistically, though, you should have 'invulnerable' characters falling to pathetically low-DC KAs rather a lot. Even when there isn't a variant or a specific 'invulnerability' build involved, though, I still haven't seen that a lot. Often, it's hand waving, the GM "doesn't bother" to roll the machine gun fire pelting the brick, because it's not meant to be a challenge, or Heros are discouraged from using KAs, or the game uses hit locations - or the guy who insists on playing the STN LOTTO loses a few times in a row, is soundly mocked, and gives up on it.

 

I'd tend to suggest many of these are what I call "operational" solutions to a rules problem; i.e. many people recognize the problem, but essentially minimize it in how they use such attacks and allow them to be used, rather than addressing them systematically. That works after a fashion, if you consider that a good solution to mechanical applications issues, which some people do.

[Opal]

Re: The great debate, this time with Java!

 

I don't consider such things fixes, no, but I have to acknowedge them. When a game is popular and has been around a while, it's players tend to learn to ignore, deny or accept it's foibles and flaws - or even have some affection for them.

[Tonio]

Re: The great debate, this time with Java!

 

I find this rather humorous, given that is exactly why I would say normal attacks are "better" - the fact that they are more predictable. Personally, I would rather be sure in the idea that my 9d6 normal attack vs 25 DEF will be doing about 7-8 STUN per attack, rather than on average my 3d6 killing attack doing 1 stun per attack (10.5 * 2.5 = 26.25).

 

As a player, this gives me the confidence that I will eventually be able to wear the target down. As a GM, it means that a lucky blow won't take out the hero, and there is also enough to fear when not getting the lucky blow.

 

The problem is that given than below a certain threshold, all results are equally undesirable, and given that there are special effects that only occur given a high enough damage, variability can become an advantage.

 

A highly consistent attack is useless if its average damage is below, or close to, defenses, since that means that it'll usually do no, or nearly no, damage. On the other hand, a highly variable attack with the same average damage will do damage more often. This is true as long as defenses as high enough to cover most of the consistent power's average damage.

 

The fact that high damages can result in stunning also favor variability, as long as the average damage isn't enough to stun the target. A highly variable attack will result in stunning more often than a consistent one.

 

Remember that the expected result is not necessarily the result of the average outcome. Consistently doing 7-8 STUN per attack, with little or no chance of stunning the target, can be substantially inferior to doing 1 STUN per attack on average, but with a nontrivial chance of stunning the target, since stunning the target can completely change the contest (it can cause the target to drop nonpersistent defenses, which might drastically raise the average STUN of the attacks, it will probably lower the DCV of the target, which will raise the effective average STUN by making more attacks hit, it reduces the damage done to the attacker by not allowing the target to retaliate, it leaves the target open to other, situational, effects, etc.).

 

Finally, remember that for a highly variable event, the "average" outcome for one event might not be as relevant as the "expected" outcome for multiple events. (That is, to more accurately judge the effect of being hit 5 times by a KA, don't calculate the average KA hit, then apply that 5 times in a row.... rather use a model to determine how likely each outcome is, and use that information instead.)

[Paragon]

Re: The great debate, this time with Java!

 

So if you use Hit locations, you slightly nerf KAs and you nerf EBs at least as much. Nope, that does not help. Well, on second thought: Since EB will already never stun (0.23% is nothing), then reducing that to 0.01% is not much of a difference, even if it is a factor of 20. It's only 0.2% really. So yeah, if you use HL, your EBs lose a bit of punch, but the KAs lose more. Still not a sound solution.

 

Well, I'm not going to speak to their use in superhero games where the very high defenses go, but I will suggest that its still a better result than the current one; honestly, having everything less likely to stun is better than having the killing attack be the attack of choice in those cases.

 

In heroic scale games, I'm rather convinced it tends to even out the field much more, however, as normal attacks actually do stun with some frequency if there isn't a big PD/attack discrepancy, and it pushes the two together considerably. I haven't seen a heroic scale game with high defense and low damage so the result may be less benign there, but see my comment above about that.

[Opal]

Re: The great debate, this time with Java!

 

NORMAL ATTACK . Stun Rule: normal damage
Plinks : 1.48 % plink ratio
Stun / atk: 11.51
Con< / atk: 0.23 %
Body / atk: 0.0

NORMAL ATTACK . Stun Rule: hit locations
Plinks : 1.48 % plink ratio
Stun / atk: 11.53
Con< / atk: 0.01 %
Body / atk: 0.0

 

This can't be right. The plink ratio has to go up, for instance, because of the 1/2 dam locations... It takes 46 raw STN damage to stun the model character & 21 to put 1 STN on him. 9d averages 31, rolling 23 or 42 on 9 dice is rare, so the clear majority of x2 results should result in stunning, and a clear majority of 1/2 dam hits should plink.

[Paragon]

Re: The great debate, this time with Java!

 

I don't consider such things fixes' date=' no, but I have to acknowedge them. When a game is popular and has been around a while, it's players tend to learn to ignore, deny or accept it's foibles and flaws - or even have some affection for them.[/quote']

 

Sure. But that doesn't mean they aren't still problem children, and as such aren't always so easily ignored. Even people who disregard them for a while can get sick of them sooner or later. We lived with the KA multiple for at least two editions of the rules, but about the middle of 4th I'd had it up to here with it, and I was regularly hearing other people commenting about it. (This, again, only refering to Champions as what "little games" we'd had locally all used hit locations as far as I can recall).

[Paragon]

Re: The great debate, this time with Java!

 

Remember that the expected result is not necessarily the result of the average outcome. Consistently doing 7-8 STUN per attack, with little or no chance of stunning the target, can be substantially inferior to doing 1 STUN per attack on average, but with a nontrivial chance of stunning the target, since stunning the target can completely change the contest (it can cause the target to drop nonpersistent defenses, which might drastically raise the average STUN of the attacks, it will probably lower the DCV of the target, which will raise the effective average STUN by making more attacks hit, it reduces the damage done to the attacker by not allowing the target to retaliate, it leaves the target open to other, situational, effects, etc.).

 

Finally, remember that for a highly variable event, the "average" outcome for one event might not be as relevant as the "expected" outcome for multiple events. (That is, to more accurately judge the effect of being hit 5 times by a KA, don't calculate the average KA hit, then apply that 5 times in a row.... rather use a model to determine how likely each outcome is, and use that information instead.)

 

These are the keys to this problem, really; the fact that a smaller amount of consistent stun is rarely as valuable as getting very little sometimes but a lot others. It becomes much more pronounced as the average damage approaches the defense value as you note, but even without that, the benefit of the intermittant gusts in taking down a target are not easily overstated; even in the case of characters with persistent defenses, being stunned potentially sets them up to get a lot of damage delivered to them that they normally wouldn't, including standard manuevers like haymakers and the like.

 

Its easy to either intuitively over or understate this effect if you've happened to hit a run in one direction or another (and if you rarely see killing attacks of a size to be noticeable used, this effect is likely to be more pronounced), but the tactical realities of combat end up tending to favor those gusts more than a simple analysis of expected stun to target will tell you.

[archermoo]

Re: The great debate, this time with Java!

 

If you haven't watched people fishing for it as a GM, and your GMs didn't have the habit, I doubt it'd be particularly noticeable, either.

 

Out of curiosity, in your own game do you also use the normal dice stun multiples from the hit location chart? If so, that'd probably tend to color your view of it too, as it ups the variance in stun over the routine somewhat, too.

 

I generally use the Hit Location chart in its entirety. Though I don't generally force people using normal attacks to use it. They can if they wish, they don't have to if they don't want to. Killing attacks have to use it, as that is where we find out how much Stun the attack does.

[archermoo]

Re: The great debate, this time with Java!

 

This can't be right. The plink ratio has to go up' date=' for instance, because of the 1/2 dam locations... It takes 46 raw STN damage to stun the model character 21 to put 1 STN on him. 9d averages 31, rolling 23 or 42 on 9 dice is rare, so the clear majority x2 result should result in stunning, and a clear majority of 1/2 dam hits should plink.[/quote']

 

The plink ratio shouldn't change at all. The normal Stun multiples only apply to the Stun after defenses, not before. The only multiple that applies before defenses is the killing one, and that is because it determines how much Stun the attack generates.

[Opal]

Re: The great debate, this time with Java!

 

On the consistency thing, consistent attacks are highly desireable when you have the advantage - as when your attack substanticially exceeds the target's defenses, but also when your defenses are good compared to your opponent's attacks, or even when the advantage is in another area, like relative CVs.

 

This comes up a lot in D&D, where some weapons or spells give more consistent results than others. In D&D, a 'normal challenge' is decidedly less powerful than the party, so consistency favors the players. In Hero, espcially in Champions, battles are often with more evenly matched opponents. In D&D, consistency is seen as highly desireable, and greater randomness less so.

 

So it's not that greater randomness is always better - it's better when you the underdog.

[Paragon]

Re: The great debate, this time with Java!

 

So it's not that greater randomness is always better - it's better when you the underdog.

 

 

Or when you need to end a fight quickly for any reason. At that point, a result that will end the fight a round later than you want is no better than one that ends it two rounds later (and may be no better than one that you lose). That isn't necessarily a common result, but it comes up.

[Paragon]

Re: The great debate, this time with Java!

 

I generally use the Hit Location chart in its entirety. Though I don't generally force people using normal attacks to use it. They can if they wish' date=' they don't have to if they don't want to. Killing attacks have to use it, as that is where we find out how much Stun the attack does. [/quote']

 

I'd still think, unless there's something I'm missing, that if you use the normal stun multiples for NPCs, and if at least some of your players use it, you're seeing more stun gusting from normal attacks than is typical for those just using vanilla, non-locational rules.

 

Of course this is a two way street, but one the whole, as noted, particularly high results are often far more noticeable and meaningful than particularly low ones, unless you have a campaign where a lot of stun is normally being delivered through defenses on typical results anyway.

[Opal]

Re: The great debate, this time with Java!

 

Or when you need to end a fight quickly for any reason.

A pressing need to end a fight quickly puts you at a serious disadvantage, yes.

[Opal]

The plink ratio shouldn't change at all. The normal Stun multiples only apply to the Stun after defenses' date=' not before. The only multiple that applies before defenses is the killing one, and that is because it determines how much Stun the attack generates.[/quote']You can tell I've never used hit locations that much. I can see how that wouldn't exactly help out normal attacks, then, and wouldn't do nearly as much as I had thought to bring the volatility of N and K damage closer together. It's also quite inconsistent, really. Aplying the KA STNx after defenses would be more consistent.

 

Even so, you should see stunning go up, not down. A 1/2 damage result will only ruin a stun result when you've rolled exceptionally high (46+), but a x2 result will turn an only slightly above average roll (33) into a stun result.

[archermoo]

Re: The great debate, this time with Java!

 

You can tell I've never used hit locations that much. I can see how that wouldn't exactly help out normal attacks' date=' then, and wouldn't do nearly as much as I had thought to bring the volatility of N and K damage closer together. It's also quite inconsistent, really. Aplying the KA STNx after defenses would be more consistent.[/quote']

 

No, it wouldn't. The StunX for KAs has to be before defenses because there isn't a Stun total for the attack to apply defenses to until after you apply it. You can apply the NStunX and BodyX after defenses because the normal Stun total and Body totals already exist and so can have defenses applied to them.

[Paragon]

Re: The great debate, this time with Java!

 

No' date=' it wouldn't. The StunX for KAs has to be before defenses because there isn't a Stun total for the attack to apply defenses to until after you apply it. You can apply the NStunX and BodyX after defenses because the normal Stun total and Body totals already exist and so can have defenses applied to them.[/quote']

 

Yeah, honestly there was no way to avoid this, as stun calculation is an intrinsic part of how normal dice are counted, where its a multiplier with killing dice.

 

That said, one of the effects this has is that the reduced stun value locations often don't mean as much if you're already only marginally leaking, either; at worst it cuts a trivial effect to even more trivial. The sweet spot of the benefit is when you're getting some amount of stun through from normal dice, but not enough to stun someone.

[Opal]

No' date=' it wouldn't. The StunX for KAs has to be before defenses because there isn't a Stun total for the attack to apply defenses to until after you apply it. You can apply the NStunX and BodyX after defenses because the normal Stun total and Body totals already exist and so can have defenses applied to them.[/quote']The KA STNx is aplied to the BOD, it would just be a matter of aplying it to the BOD after defenses.

 

That would be more consistent with the way BOD and N-STN multipliers on the hit location chart work: after defense.

 

For instance, say you're hit in the head by an attack that barely hurts you - you have 12 PD, 6r, and get hit for 7 BOD, you'd take 1 BOD (doubled to 2) and 5 STN. Just like if you took a normal attack that barey penetrated your defense, getting only say 3 STN past you'd only take 6 STN, not 18. OTOH, the way it works now, you'd take 35 STN against your total defense and take 23 STN, for a 1 BOD knick.

 

The KA and N dice conventions obviously aren't consistent with eachother, but the way they aply multipliers for hit locations could be, if you simply aplied all the modifiers after defenses. That's consistent.

[Paragon]

Re: The great debate, this time with Java!

 

The KA STNx is aplied to the BOD, it would just be a matter of aplying it to the BOD after defenses.

 

 

That would have the effect of making resistant defense absolute against killing, and mean that no stun was possible from it unless you did Body.

[Silbeg]

Re: The great debate, this time with Java!

 

NORMAL ATTACK . Stun Rule: normal damage
Plinks : 1.48 % plink ratio
Stun / atk: 11.51
Con< / atk: 0.23 %
Body / atk: 0.0

KILLING ATTACK . Stun Rule: 1d6 -1
Plinks : 44.29 % plink ratio
Stun / atk: 11.57
Con< / atk: 17.43 %
Body / atk: 1.45[/quote]

Without making me dig through your code, could you tell me how many die rolls you were making per test?

Out of curiosity (and desire to avoid the crap I am supposed to be doing at work) I wrote a little tool to do the same.

My results were very similar...

[code]
Normal Results:
Total Attacks: 100000000
Average Damage: 11.52
plink %: 1.49
stunned %: 0.24
---------
Killing Results:
Total Attacks: 100000000
Average Damage: 11.57
plink %: 44.29
stunned %: 17.44

 

As an FYI - I was using 9d6/3d6, and standard STUN multiplier.

 

I was interested, since our numbers were very close, but not exact. I think rounding modes would also make a difference...

 

Example of my rounding...

 

 

  public BigDecimal getAverageStun () {
       double averageStun = totalDamage / numberOfAttacks;
       BigDecimal value = new BigDecimal( averageStun );
       value = value.setScale( 2, RoundingMode.HALF_UP );
       return value;
   }

 

This gives 2 digits of precision, with a value of .005 rounded up to .01.

 

This is mostly curiousity (and some laziness?)

 

I could bundle my java source up, if you are interested.

[archermoo]

Re: The great debate, this time with Java!

 

The KA STNx is aplied to the BOD, it would just be a matter of aplying it to the BOD after defenses.

 

That would be more consistent with the way BOD and N-STN multipliers on the hit location chart work: after defense.

 

Changing it to multiplying Body that gets through defenses is a massive change to how KAs work, and does nothing to make anything more "consistant".

 

Again, the StunX for KAs is before defenses because it is how you figure out what the baseline Stux for the attack is. It isn't a multiple to an existing amount of damage to reflect that area being more or less sensative. It is the mechanical way to determine how much Stun is done by the attack. It is doing something different than the other multiples do, so changing the nature of Killing Attacks doesn't make it consistant, it just changes how KAs work.

 

If you wish to drastically change the way in which Killing Attacks work in your campaigns, house rule it to your hearts content. But the result isn't any more consistant with anything.

[Opal]

That would have the effect of making resistant defense absolute against killing' date=' and mean that no stun was possible from it unless you did Body.[/quote']Correct. Which would certainly solve the problem of KAs being too effective against high-resistant defense characters, as well as being more consistent with the way other multiples due to hit locations work.

 

I know it wouldn't model too well the way 'real armor,' works, but that could be addressed as part of the real armor limitation, in non-supers games.

 

 

Of course, things like bullet proof vests and chainmail rarely prevent bullets and swords from doing damage that takes a little more than a few minutes to recover from, either, so it's not all bad, linking STN to BOD. In my campaign, I require most 'real armors' to have a limitation (I call it Flawed -1/2) that makes all attacks act like penetrating attacks against them. It's usually trivial with regards to normal atacks, but it means that KAs /do/ hurt through kevlar.

 

There are probably better ways of handling KAs vs N attacks in modeling such things, but they'd require something like changing how KAs are rolled based on the defesnses of the target.... I've considered some such alternatives, but never implemented them.

 

 

If you wish to drastically change the way in which Killing Attacks work in your campaigns' date=' house rule it to your hearts content. But the result isn't any more consistant with anything.[/quote']As I mentioned before, I have done so. But, honestly, it does make KAs more consistant with other attack powers. It makes them more internally consistant, in that thier primary practical use becomes killilng low-restiant-defense targets, instead of stunning invulnerable bricks. It makes the cost of defending against a KA more consistent with the cost of defending agaisnt other attacks that go against other-than-normal defenses. And, it turns out, makes the aplication of hit location modifiers more consistant, as well.

 

Where it falls down, IMX is modeling realistic, normal armor (see above).

[archermoo]

Re: The great debate, this time with Java!

 

 

As I mentioned before, I have done so. But, honestly, it does make KAs more consistant with other attack powers. It makes them more internally consistant, in that thier primary practical use becomes killilng low-restiant-defense targets, instead of stunning invulnerable bricks. It makes the cost of defending against a KA more consistent with the cost of defending agaisnt other attacks that go against other-than-normal defenses. And, it turns out, makes the aplication of hit location modifiers more consistant, as well.

 

Where it falls down, IMX is modeling realistic, normal armor (see above).

 

I have no doubt that your modified KA is internally consistant with your vision of what your modified KA should be. That that bears very little on the actual unmodified Hero rules however. And doesn't make anything more consistant.

[Silbeg]

Re: The great debate, this time with Java!

 

Now, just to make things interesting, I changed the parameters to something that is more likely in my Champions game.

I used the following:

DC = 12 (12d6 normal, 4d6 killing)

CON= 25

DEF = 25

Since I don't really care that much about the BODY damage, I didn't keep track.

But...

 

Running Killing (4d6)
Running Normal (12d6)
Normal Results:
Total Attacks: 100000000
Average Damage: 17.00
plink %: 0.21
stunned %: 7.60
---------
Killing Results:
Total Attacks: 100000000
Average Damage: 16.29
plink %: 40.09
stunned %: 29.07

 

So, in this case, you are 200X more likely to "plink" with a killing attack, and only a little less than 4X more likely to STUN when when killing. Now, that seems a pretty high likelihood of stunning... but we are talking pretty much stun or plink here, since you are about 30% likely to do non-stunning damage.

 

Again, to me, I would prefer to do the 17 damage each attack (on average). For a typical target, that would mean I get a KO in 3 shots, while in 3 shots I might stun the person, but won't be as likely to achieve a KO.

 

That makes me think... maybe I should modify to collect that stat as well... number of attacks required to KO (assuming 40 STUN?).

 

While I can imagine the result, I think the statistics would be the most telling.

 

Maybe tomorrow... now it is time for V-Day

[Opal]

Re: The great debate, this time with Java!

 

Which of these things is not like the others?

 

Apts of defense to 'bounce' 30 Apts of attack power:

 

Ego attack: 18

 

Flash: 12

 

Killing Attack: 66

 

Drain: 18

 

AVLD: 13-19

 

 

Would you say that the amount you have to invest in an apropriate defense is consistent from one attack power to another?