Deceptive Yak Shaving – Deceptive Attack Redux
A thread on the forums made me break out Excel again so I can look at the “optimum” one-one-one level of deceptive attack for all combinations of DA and defense level.
The answer is darn complicated, turns out.
Here’s an example of what I mean:
This is the example Cadmus faces when he’s got his Righteous Fury on and is swinging with Axe/Mace-24. You can see that how much Deceptive Attack he throws – after accounting for lots of other choices like targeting foe’s important bits – depends hugely on his enemy’s skill level. Note that the formatting lies a bit – if your net skill is better than 16 and your foe sucks, by all means drop it to 16. You have nothing to lose.
That “go to 16” plan looks good (within 5% of the best choice, or itself the best choice) up to the point where my foe is sporting a defense roll of about 16 or more. Of course, that’s a shield, a retreat, and Dodge-11 (+2 shield and +3 retreating dodge) or Parry-13 (+3 for a medium shield and a +1 retreat). That’s Skill-20 for my foe, so he’s a real champ – Cadmus’ level.
Point is, “how much do I DA” is actually a non-trivial question in many cases. Well, not for the case of skill 14 or less, for which the answer is simple: don’t do it.
Note that I included my spreadsheet for download if you want to poke.
Now, my spreadsheet (yes, I have one too) doesn't include anything about criticals, but it does say that if you have a net skill of 14, you can still afford to drop to 12 if your opponent's defense is 10 or more, 13 can afford it for a skill of 12 or more, and even a skill of 12 can get a benefit from one level of DA against Def 13 or higher.
Granted, those benefits are marginal at best, one and two percentage points of boost.
The criticals are pretty important, though, since they set a floor on how bad you can suck with a foe with great defenses of about 10% (9% for your crit, and the extra comes from the low odds he'll fail or crit-fail his roll).
But the real key to "why not eke out that marginal 1-4% benefit" is in the good stuff that a crit buys you (more damage, foe stuns himself, unreadies his weapon) and the reduced chance of you rolling your own crit fail (it's a low chance, but with skill of 16+ it's a 3.5x reduction in the frequency of that outcome).
And as Peter noted in that other thread, a poorly timed critical (or a well-timed one, depending on point of view) can turn the tide of battles, since they allow or prevent concentration of force.
The rather epic battle in http://gamingballistic.blogspot.com/2014/04/gurps-jade-regent-crikey-yetis.html the Jade Regent session started to turn when a few of us critically hit a pair of Skill-30 foes. After that, no one could touch us from a mundane standpoint, but 'till then, we were facing real threats. That sort of knife edge is why Peter's players avoid dropping below 16, I think.
"That sort of knife edge is why Peter's players avoid dropping below 16, I think."
Pretty much.
Take the case of Vryce – he's heavily armored, has very high defenses, high damage, lots of HP, and a two-hex reach weapon with Sacrificial Parry . . . and said weapon is magical and sometimes the only thing he has to defeat his foe. The consequences of rolling a 17 or an 18 and having that be a critical failure is that he goes from a protective Parry unbrella that is also dealing damage (or threatening to do so via attacks) to a non-factor in one bad roll. Closing down the margin on that roll is more important than taking out, what, 5% more combatants over the same given amount of time?
Plus as you note, timely criticals can turn a fight.
And, as we've discussed offline many times, the "confused melee with many opponents" is the default in my campaign. The consequences there of critical hits and especially critical failures are quite high, and thus (most of) my players take them into account as a matter of course.
When I was shaving this Yak, I noticed that for any skill/advantage/weapon situation where you can Rapid Strike, RS is probably a better solution than a comparable spend on deceptive attack.
The main exception is when enemy defense is close to 16, but it is really hard to know that as a player. Eyeballing your table above this still applies.
Additionally, for enemies where a hit is not the same as taking them out (that is to say, the ones where optimization is important), hit locations are also probably better than DA. So in general I recommend a skull shot or two shots at the leg instead of DA 8 for skill in the 24ish range*, unless you have reason to believe enemy defense rolls are 15-18.
(*) If you are a WM, consider three or even four shots. If you allow extra effort for RS, or have extra attacks, Just Go Nuts. Those 10% crit chances for 16 + effective skill really add up, and multiple crits in the same round on the same enemy are Fun.
Agreed. The "Attacker Skill-20" column should be read as "after you take things off the top like hit location and rapid strike penalties." It's also after adding maneuver-based modifiers like Committed Attack. As I noted in the post "A Surplus of Awesome" from way back when, there's lots and lots to do, in 4e, with high skill. But once you do the things you want to do in terms of deciding how many blows to throw, and where to throw them, the question of how deceptively to launch each one comes up, and that's where the sheet comes in.
Mathematically, anyway. Not all of these things are easily subsumed into math.
Your spreadsheet seems to have very marginally different results from what I've gotten, and may have a rounding error. Looking at the case of 24 vs 11 with deceptive attack 4 (so 16 vs 7), we should have:
20/216: critical hit (9.259%)
192/216 * 181/216: normal hit, defense fails (74.486%)
Total 83.745%.
I suspect my original sheet (I have another one I can't upload until I got home) is off by a bit because of my P(Crit) + (1-P(Crit))*P(Hit)*P(Fail Defense) instead of P(Crit) + [P(HIt)-P(Crit)] * P(Fail Defense).
The results aren't that different. I also have a new table that identifies options that are within 5% of the "best" option for any given row. THAT result was much more conclusive and much less complicated: the best overall strategy for skill 16+ is to reduce skill to 16. This is within five percent of the best option for nearly all cases (maybe three rows) when you don't know your foe's defenses.
The conservative method of reducing skill to 16 can have up to an 18% worse hit chance than a basic attack (when it reduces opponent skill to 13; 16 vs 13 is 23.7%, 12 vs 11 is 28.9%), but there are two additional factors favoring only reducing to 16:
1) A parried or blocked hit consumes an active defense (resulting in multi-parry penalties)
2) Critical hits have somewhat greater value than unparried normal hits — 62.5% of them have some sort of increased effect.