Question about the maths behind attack roll and defense rolls

[SolemnGolem]

So I was leafing through the GURPS 4th ed. rules and looking at the opposed roll concept, which I liked. One thing stuck out at me.

Attack and defense are not opposed rolls. They're close, but the important issue of margin of success is missing. The way it is, Attacker rolls against her skill - failure means it fizzles, but success merely means Defender rolls against his skill. Defender does not apply a penalty to this roll based on Attacker's roll.

From a mathematical standpoint, what's the effect of this? If the designers had decided to make it an opposed roll, and applied Attacker's margin of success as a penalty to Defender's defense roll, how would the game be different?

The only thing I can think of is Defenders would fail more often - pretty much on par with the probability of standard opposed rolls as they currently stand. Did the designers remove the margin of success penalty to make the game less lethal? Has anybody applied the margin of success as a penalty, and how has that made your game run?

[laguna]

what it means is that if someone has 28 sword skill, they have a 95% chance to parry and can only be hit by critical strikes, if i understand correctly.

[gjc8]

Originally Posted by SolemnGolem

From a mathematical standpoint, what's the effect of this? If the designers had decided to make it an opposed roll, and applied Attacker's margin of success as a penalty to Defender's defense roll, how would the game be different?

The current system has the deceptive attack mechanic to make it effectively into a contest. The difference is that the attack must commit to a penalty before hand. This is a significant disadvantage.

It also gives an relative advantage to attackers who know what the defender's skill is.

[laguna]

Originally Posted by gjc8

The current system has the deceptive attack mechanic to make it effectively into a contest. The difference is that the attack must commit to a penalty before hand. This is a significant disadvantage.

It also gives an relative advantage to attackers who know what the defender's skill is.

hadn't noticed that.

huh.

[gjc8]

Originally Posted by laguna

hadn't noticed that.

huh.

It's a key feature of the system. 3e lacked such a system by default, and suffered exactly the problem you pointed out.

The books, unfortunately, don't really explain how important it is.

[laguna]

Originally Posted by gjc8

It's a key feature of the system. 3e lacked such a system by default, and suffered exactly the problem you pointed out.

The books, unfortunately, don't really explain how important it is.

yeah, it had been my intention to expand the crit numbers beyond 6 (for skill 16) to 7 and beyond so that two master swordsmen would chop each other to pieces with crits instead of endless parries.

now i'm thinking that won't be necessary.

[Landwalker]

Originally Posted by laguna

yeah, it had been my intention to expand the crit numbers beyond 6 (for skill 16) to 7 and beyond so that two master swordsmen would chop each other to pieces with crits instead of endless parries.

now i'm thinking that won't be necessary.

Why would you expect two master swordsmen to rapidly dismember each other instead of to have a long, drawn-out duel?

[DukeofDellot]

When I first saw this, I thought of how a duel between two highly skilled individuals might last much longer than that between two poorly skilled individuals... in fact I believe I made a hit chance chart... I may still have it. In the long run, it didn't matter... deceptive attacks and feints put that twelve minute's effort down the tube... but I believe it's called... um... er... bagu...

Multipicitive Percentages?

Take the chance you have to make your attack roll, and multiply it by the chance you have to make your active defense. If I had my books on me, I could put a page number on where you'd find the odds, it's at the begining of the "Skills" chapter in the Basic Set Characters.

Apparetly I still have the chart though... if both characters have a skill 15, no enhacements to parry, and the attack goes unmodified... you have a 50% chance to hit, but if both characters have equal skill and combat reflexes (granting a +1 to active defense rolls), you have roughly a one in three chance to hit an unmodified attack until you get into the 18+ skill range.

Eh... at least I kind of thought that was interesting.

Edit: Come to think of it, Deceptive attacks bring the 1/3 chance to hit on equal skills up and beyond the 18+ skill range, if that's all you're doing with your attack.

[Landwalker]

Originally Posted by DukeofDellot

Take the chance you have to make your attack roll, and multiply it by the chance you have to make your active defense. If I had my books on me, I could put a page number on where you'd find the odds, it's at the begining of the "Skills" chapter in the Basic Set Characters.

Not quite: It would be the chance of a successful attack roll multiplied by the chance of a failed defense roll.

So assuming no other bonuses, no fancy maneuvers, and combatants with Skill 16 (Parry 11), each attack has a success rate of roughly 36.7875%.

If the attacker takes a -2/-1 Deceptive Attack (giving target rolls of 14 and 10), that improves to about 45.35%

A -4/-2 Deceptive Attack gives a slight additional improvement, to about 46.3%.

Edit: Incidentally, if you decrease each character's base skill by 1 (to 15 each), the odds of that unmodified attack hitting jump up to 47.7%, and a -2/-1 Deceptive Attack gives you 52.4%.

[SuedodeuS]

The concept is that the attack roll is answering the yes/no question of "Did you swing well enough to connect?" while the defense roll is answering the yes/no question of "Did you get out of the way/interpose an object in time?" The reason Quick Contest mechanics aren't used is because of choice. I'm not particularly good at combat - I might have Karate Sport at DX. Against someone who's decent, I can throw attacks that should connect all day, only to have them Parried or Dodged. I can increase my chances of scoring a hit by getting tricky - a quick fake-out before an attack, a particularly rapid attack, trying to hit that little opening I see, etc. In all cases wherein I do this, however, there's a trade-off - I have a higher chance (and I've actually noticed this) of missing outright, without need for my opponent to do anything to avoid the hit. This is what GURPS calls a Deceptive Attack*, which imposes -2 to hit for -1 to the enemy's defense. From the other side, the situation is fairly comparable - I can Parry or Dodge (retreating in both cases) their normal attacks fairly well, and they will frequently do a fake-out or quick attack in order to connect. The important thing here is that there is a risk, and the fighter must choose to engage in this risk in order to benefit from it. When you make an attack, you (in general) make the attack you were intending - you don't mess up and suddenly make it better.

That said, including Quick Contest like mechanics won't break the game, although they will result in more hits. The element of risk on the part of the attacker is lessened, as he needn't sacrifice his chances of hitting to decrease the chances of a successful defense. Do note, however, that Parry is based on half one's skill level - so it would be appropriate to half the attacker's Margin of Success to determine the penalty. If you don't like doing this, you could adopt the rules from T-bone's FEND (Fully ENabled Defenses) system, where Parry is Skill, rather than (Skill/2)+3, Dodge is Basic Move*2 (instead of Basic Move+3), defenses take a general -4 penalty, and penalties from Feints are doubled. In that, Deceptive Attacks trade off on a 1-for-1 basis - using Quick Contest mechanics, this simply means you are automatically going as deceptive as possible and still landing a hit.


*At my skill level, I'm probably doing something more akin to Telegraphic Attack normally, and then simply being less telegraphed when I do what I'm calling a Deceptive Attack. The math works out the same either way, however.