Evaluate useless? (from Why Wait)

[Lupo]

Originally Posted by Kromm

We put Evaluate into 4e specifically to differentiate aggressors (who start combat after three Evaluates, for +3) from their victims (who enter combat after a series of Do Nothings); it's worded in terms of a second-by-second combat maneuver only for consistency. We didn't seriously think that fighters would often select it (or Do Nothing!) in a raging fight.

I won! :P The Great Kromm confirmed that Evaluate is not even meant to be comparable with other maneuvers in a "raging fight". I won :P

Joking aside, I think this thread has been very interesting, both for people who like Evaluate and for people who think it is "too weak".
The thread provided many situations where Evaluate *is* indeed useful (surprising some of us "Evaluate-haters"), and debunked some persistent myths about Evaluate being generically useful to improve your chances of hitting (surprising some of the "Evaluate-lovers")
I will try to recap what has been said, this could be useful as a memorandum for tactically-inclined players:

A brief essay on Evaluate, by the Hivemind

Evaluate seems to always be inefficient when you are within Step + Reach of a ready opponent, unless there are special circumstances. Attacking two times gives a higher chance of landing a blow than attacking once with +1 to skill .
There are, however, many special circumstances and situations where Evaluate becomes a good tactical choice, or the best possible option.

1) when your opponent does not see you (e.g. ambushes) or he is not aware a fight is about to start (e.g., your character might use Evaluate on a particularly dangerous potential opponent in a negotiation, just in case the negotiations go sour).

2) If your weapon is unbalanced, or will become unready after you strike, or otherwise when attacking costs you resources. For instance if attacking costs you a fatigue point each time you attack (such as from an imbuement skill) and you need your attacks to land because of this evaluate becomes a better an increasingly better option to take as your resources get lower and lower.

3) when you absolutely need to make a difficult shot

4) while advancing toward an opponent who is distant(like 4-5 yards), as you can evaluate as you walk forward. If you don't want to Move and attack (which is risky), Evaluate can be used at some range particularly with a long reach weapon. At least in the first turns this might be handy.

5) If your opponent is skilled in Counter-Attack or makes a lot of ripostes, attacking every turn is a bad idea.

6) Evaluate also hands you nice defensive bonuses against feints.

7) several techniques are used when the opponent has done a parry on your attack (like a Judo throw): Evaluate, then make a very deceptive attack to avoid the parry and the risk to finish prone or worse.

8) When your misses become dangerous. Fairly rare but occasionally an important consideration when the opponent is using something or someone to shield himself from you and if you hit on a miss or successful block by him its a bad thing from your sides prospective. For example a demon using a child as a living shield against your character and any successful block by the demon means that you hit the child.

9) When your opponent will let you have 1 free hit before he [does something]. Once you attack you have to deal with the event triggered, so it's better to Evaluate first, to maximize chances.

10) if, for any tactical reason, you want to move backward, using Retreat, you must not attack to avoid that your opponent does the same thing; then, after several moves, you have a chance to land an accurate attack;

11) if your opponent has a better reach than you and stays out of your reach, you can use Evaluate to prepare an All-Out Attack that may break this situation.

==

On the other hand, Evaluate is NOT a good tactical choice:
- when you just want to hit your opponent. Repeated attacks are quicker and more reliable than Evaluating + Attacking.
- when you just want to overcome your opponent's defenses. Feints and multiple attacks are (usually) more efficient in that regard.
- when you fear your opponent's attacks, and you don't want to expose yourself to more of them by "slowing down" the fight Evaluating.

[vicky_molokh]

Originally Posted by Kromm

My players use Evaluate as written. It's a special-purpose maneuver whose niche is "making use of a turn on which an attack wouldn't be the ideal choice," such as when waiting for allies to get into position before rushing someone. It mostly sees use immediately before a sudden attack that initiates combat; in effect, it lets the side that initiates action do so at +3. It isn't as generally useful as Attack or Move . . . but then, all maneuvers aren't of equal value. Look at Do Nothing! We put Evaluate into 4e specifically to differentiate aggressors (who start combat after three Evaluates, for +3) from their victims (who enter combat after a series of Do Nothings); it's worded in terms of a second-by-second combat maneuver only for consistency. We didn't seriously think that fighters would often select it (or Do Nothing!) in a raging fight.

It seems like a niche manoeuvre for when you're facing your potential opponent (face to face), because backstabbing is likely better done with Telegraphic Attack. Of course, there are Perks which change the way Evaluate works, which make it more useful.

[Siddhartha]

Originally Posted by Wicked Lurker

So far I have not yet found a use for evaluate and actually think it's useless in most situations.

Consider the the alternatives:
A Skill-12 warrior attacks some adversary.

1) choose evaluate, attack next round: one attack at skill 13. probability of success 0.84
2) choose attack and also attack next round: 1-((1-0.74)^2). probability of success 0.93

There's something wrong about it:

If a skill-12 warrior chooses to attack and also attack next turn, the probability of success is 0.74 * 0.74 = 0.55

So, we have two different situations here:

1) you Evaluate and increase your probability to make one successful attack about 10% ( (skill 13=83,9%) - (skill 12=74,1%) )
2) you try to make two successful attacks (skill-12), but the probability of this happen is 55%.



I think Evaluate isn't useless at all !

[Ze'Manel Cunha]

Originally Posted by Siddhartha

There's something wrong about it:

If a skill-12 warrior chooses to attack and also attack next turn, the probability of success is 0.74 * 0.74 = 0.55

So, we have two different situations here:

1) you Evaluate and increase your probability to make one successful attack about 10% ( (skill 13=83,9%) - (skill 12=74,1%) )
2) you try to make two successful attacks (skill-12), but the probability of this happen is 55%.



I think Evaluate isn't useless at all !

Bad math on your part, no cookie.

[aesir23]

Originally Posted by Siddhartha

There's something wrong about it:

If a skill-12 warrior chooses to attack and also attack next turn, the probability of success is 0.74 * 0.74 = 0.55

So, we have two different situations here:

1) you Evaluate and increase your probability to make one successful attack about 10% ( (skill 13=83,9%) - (skill 12=74,1%) )
2) you try to make two successful attacks (skill-12), but the probability of this happen is 55%.



I think Evaluate isn't useless at all !

No fair comparing the chance to hit at-least-once with the chance to hit twice. Apples and oranges.

[Dinadon]

Originally Posted by Siddhartha

There's something wrong about it:

If a skill-12 warrior chooses to attack and also attack next turn, the probability of success is 0.74 * 0.74 = 0.55

Mathematics fail. You're only counting the situation where both attacks hit.
The comparison is getting one successful attacks in two turns, thus you need to count the situations where only one of those attacks hits. Which was done.

[RyanW]

Originally Posted by Siddhartha

If a skill-12 warrior chooses to attack and also attack next turn, the probability of success is 0.74 * 0.74 = 0.55

That would be the probability of two successes.

The probability of at least one success is equivalent to 1-(probability of two failures). In other words: 1-(1-.74)*(1-.74)=0.93.

That said, there are plenty of times you want to increase your chances of not missing, rather than just the chances of hitting. When you might hit the wrong target, for example, or when you want to limit the number of attacks made. Innate Attack (Melee, Costs FP) perhaps?

[sir_pudding]

Originally Posted by RyanW

That said, there are plenty of times you want to increase your chances of not missing, rather than just the chances of hitting. When you might hit the wrong target, for example, or when you want to limit the number of attacks made. Innate Attack (Melee, Costs FP) perhaps?

A situation that has already been discussed and nicely summarized in Lupo's last post.

[Siddhartha]

Originally Posted by Ze'Manel Cunha

Bad math on your part, no cookie.

Originally Posted by aesir23

No fair comparing the chance to hit at-least-once with the chance to hit twice. Apples and oranges.

Originally Posted by Dinadon

Mathematics fail. You're only counting the situation where both attacks hit.
The comparison is getting one successful attacks in two turns, thus you need to count the situations where only one of those attacks hits. Which was done.

My bad ! =)

So, to skill-12:

X) 1 hit/1 miss = 0.74 * (1-0.74) = 0.1924
Y) 1 miss/1 hit = (1-0.74) * 0.74 = 0.1924

And the probability of one of the attacks hits is X + Y = 0.3848


Comparing with the "Evaluate alternative" (84% to hit), you have more than twice the chance to hit.

Originally Posted by RyanW

That would be the probability of two successes.

The probability of at least one success is equivalent to 1-(probability of two failures). In other words: 1-(1-.74)*(1-.74)=0.93.

Wrong! When you do this, you're also including the probability of two successes...

2 hits = 0.74 * 0.74 = 0,5476
1 hit/1 miss (or vice versa) = 0.3848 --> calculated above
2 misses = (1-.74) * (1-.74) = 0,0676


You should take your 0.93 (0.9324 to be exact) and subtract the 2-hits probability (0.5476)

And guess what? We have 0.3848 precisely.

[Ze'Manel Cunha]

Originally Posted by Siddhartha

Wrong! When you do this, you're also including the probability of two successes...

Bad Siddhartha, still no cookie for you.

You absolutely need to include the probability of two successes, and since it's much better to hit twice, than to merely hit once, that makes your posts all the stranger.