Optimum Level of Deceptive Attack

[lexington]

What do you mean by "percent benefit"?

[dublindog]

Originally Posted by lexington

What do you mean by "percent benefit"?

If your initial odds of a landed blow (attack succeeds, defense fails) are 50% [Skill 15 against Defense 10], you look on the first chart and see that you can take 1 Level of DA (-2 skill, -1 defense) and improve your odds of a landed blow to 53.07%. This is a percent benefit of 3.07%.

So, I'm not using initial probability as the denominator. If you wanted to in this case, it's a 6.14% improvement. But since the actual important number is the final odds of landing a blow, I'm using straight probability out of 100.

So, you can ADD the "Percent Benefit" to the initial chance of landing a blow to find the resulting optimized chance using DA.

[kure]

I DO like stuff like this, but i can see some flaws on this table.

if I have skill like 20, why ever should I get only deceptive attack 1?

result will be Attack 18 and defense penalty -1

when all rolls 17+ is fail anyway, it have no benefit to have net skill 18.

If your initial odds of a landed blow (attack succeeds, defense fails) are 50% [Skill 15 against Defense 10]

is it?
defense roll 10- alone have 50:50 or 50% chance,
but there is also chance of failing 15- attack roll which affect result.

[dublindog]

Originally Posted by kure

I DO like stuff like this, but i can see some flaws on this table.

I'm not perfect, so there could be an error, but you're going to have to be more specific on what flaws you see. You might be mistaking your own misconceptions with flaws in the table.

When I started this task I neglected critical hits and boundary conditions not thinking they'd matter. But they do. I had to redo the chart recognizing that effective Defense 4 is the lowest effective (not 3), that 16 was the highest effective Defense and Skill. The odds don't change below 4 and above 16.

if I have skill like 20, why ever should I get only deceptive attack 1?

result will be Attack 18 and defense penalty -1

Notice the Defense level. Defense 5, becomes Defense 4. Defense 4 is the lowest effective Defense level. You could hypothetically lower this further, but it will remain the same probability of a successful defense, i.e. a critical roll 3 or 4.

when all rolls 17+ is fail anyway, it have no benefit to have net skill 18.

The skill level isn't the deciding factor here, it's the defense level. AND, like several spots on the table, you can take another level of DA and get the exact same P[Landed Blow].

P[18,4] = P[16,3] = 94.03%

The way I programmed the chart, it lists the most conservative level of DA that still gives the optimal result.

is it?
defense roll 10- alone have 50:50 or 50% chance,
but there is also chance of failing 15- attack roll which affect result.

Yes, as I stated before, Attack Skill 15 vs. Defense Skill 10 is a 50% chance of a landed blow. You are neglecting criticals. The defender does not defend successfully 50% of the time. He defends successfully 50% of the time that the attacker does not get a critical success. When the attacker gets a critical success (Roll of 3,4,5 on a skill of 15), he gets NO defense. So over all his odds of defending are less than 50%, because 4.63% of the time the attacker will get a 3,4, or 5.

P[Landed Blow] = P[Attacker Criticals] + P[Attacker Succeeds w/o Critical] * P[Target Defends]

P[15,10] = 4.63% + 90.74% * 50%
P[15,10] = 50%

Can you see how this works?

[kure]

Originally Posted by dublindog

I'm not perfect, so there could be an error, but you're going to have to be more specific on what flaws you see.
...

Can you see how this works?

My bad, i just wrote faster than make math myself. DA is less intuitive than i realize first time.
Second case i just forget Criticals, but result of neat 50 did look suspicious :)

Anyway Im sorry :)

[Bruno]

I suspect everyone who tries this sort of analysis forgets criticals in their first draft.

I sure did :D

Those 0 regions on the defense chart are prime areas for switching to Dual Weapon Attack or Rapid Striking (or if you're crazy, All Out Attack Dual) to try and improve your chance for a critical hit, thereby bypassing the defenders amazing defenses completely.

They're even better regions for you to start to consider completely changing your tactics :D

[vicky_molokh]

Originally Posted by Bruno

I suspect everyone who tries this sort of analysis forgets criticals in their first draft.

I sure did :D

Those 0 regions on the defense chart are prime areas for switching to Dual Weapon Attack or Rapid Striking (or if you're crazy, All Out Attack Dual) to try and improve your chance for a critical hit, thereby bypassing the defenders amazing defenses completely.

They're even better regions for you to start to consider completely changing your tactics :D

And then there's the diminishing Parry/Dodge score...

Really, analysis beyond the 1 attack / 1 defense is too complicated to use in real time without Intuitive Mathematician or (Eidetic Memory + Lightning Calculator).

[Lupo]

A really interesting work...

Allow me to make some observations:

1) often, a Deceptive Attack will grant a very minor benefit. It is somewhat less useful than I thought. E.g., for attack skill in the 15-17 range, a DA improves your chances to hit by +5%/+10% AT MOST.

2) if the enemy best defense is not Dodge, the DA benefit will be even lower than the tables indicate. That's because forcing your enemy to use up his Parries/Block will penalize his further defenses; so having your attacks Blocked or Parried is often better than risking to fail them outright with a DA. (because other fighters could, and will, attack your enemy forcing him to use his lower dodge).

3) For very good defenses VS not-so-good attacks, the "best DA option" is often not intuitive at all, and it changes weirdly. E.g., if you are facing an enemy with an effective Defense of 16, and your attack skill is 16, your best option is an ordinary (not Deceptive) attack. But if your attack skill is either 15 or 17, your best option becomes a -6/-3 Deceptive attack!

[Ulzgoroth]

Originally Posted by Lupo

3) For very good defenses VS not-so-good attacks, the "best DA option" is often not intuitive at all, and it changes weirdly. E.g., if you are facing an enemy with an effective Defense of 16, and your attack skill is 16, your best option is an ordinary (not Deceptive) attack. But if your attack skill is either 15 or 17, your best option becomes a -6/-3 Deceptive attack!

...How do you figure?

With skill 15, -6/-3 isn't even legal.

With skill 17, -6/-3 by my figuring gives a slightly lower total chance to hit and a vastly lower chance of criticals than no DA.
EDIT: I fumbled the math. -6/-3 is (a little) better real hit chance.

[NineDaysDead]

Deceptive Attack speadsheet by seasong:

Originally Posted by seasong

Spreadsheet love:
Deceptive Attack

Enter any two attack and defense rolls and it will calculate the probabilities for up to twenty levels of Deceptive Attack (my science fiction campaign sometimes hits 30+ effective skill rolls), and also explicitly point out the optimal number of Deceptive Attack levels.

Most of the actual process is in the hidden columns, if you want to look at the specifics - I hide them to keep things visually uncomplicated.