[gen.nehring]

Hey, i was wondering if somebody who likes figuring out statistics would like doing this.

STATISTICALLY, what has a better survivability against standard bolter fire, the T4 but only 4+ save scout, or the T3, but 3+ armoured battle sister?

 

[DominionWeaponsGroup]

it takes 6 bolter shots to kill a scout (2/3)x(1/2)x(1/2)=1/6
as opposed to 7ish to kill a sister (2/3)x(2/3)x(1/3)=4/27

and no, i didnt really like doing this

 

[Bean]

Yes, what he said. 4/27 chance of killing a sister, 1/6 chance of killing a scout. It's slightly easier to kill scouts.

 

[ForgedInTheFurnaceOfWar]

Thats why power armour is so feared.

 

[Sabe]

I actually ran a statistic analysis on the survivability of like 30 or 40 different troop types and found that overall(though not exclusively) your models survivability depends on:

1. Save(most important)
2. Toughness
3. Wounds(least importantly)

In some cases the models toughness is more important as it makes them immune to certain Strengths. Though many times the toughness is more important anyway because an attack ignores your save. Very rarely are wounds of any importance passed 2nd wound. For example a Necron Lord(T5,W3) is more survivable than a Shas'o(T4,W4) even before you factor in WBB for the necron lord. This is also calculating that either neither of them have the 4+ inv save available to there army or both do. The only exception to this is when the str. of the attack becomes strong enough to wound either on a 2+, in which case the extra wound is better. Mind you...even then the Lord has WBB.

Overall however you'll find that your survivability is based on those three things in that order.

 

[Bean]

The lord can still be instagibbed by S:10 weapons (denying him his WBB roll).

That being said, you're only sort of correct.

It may be the case in most situations encountered during the game that a better armour, lower toughness unit is more survivable than a worse armour high toughness unit, but don't let that mislead you.

In actuality, each part figures into the equation with exactly the same weight.

Take the instance of guard and orcs being hit by lasgun fire.

Guard take 1/2 of the hits in wounds and 2/3 of the wounds in kills, or 1/3 of the hits in kills.

The guard have a better save, but they die more to lasgun fire.

What you'll actually find is that the model type whose positive aspect (that is, its stat that is better than the one you're comparing it to) produces target numbers farthest from the average roll will perform the best (presume the variation between total relavent stats is zero, and none of the game-imposed limits are reached--such as the instance you mentioned where both get wounded on 2+ despite the difference in toughness)

To illustrate, consider the sisters and the scouts. The scouts were better than the sisters in toughness, and a 4+ (1/2 chance) was needed to wound them. The sisters were better than the scouts in armour, and there was only a 2/3 chance of a wound getting through. The 2/3 was farther from the mean than the 1/2, and the sisters perform better.

You'll see the same thing is true in the case of the guard.

 

[Sabe]

My formula says that the sisters are better too...

And I made it very clear that that was an Overall statement.

And as for your example on the Necron lord the Shas'o gets instakilled too...and very few lords don't have a res orb for "actualities" sake. Like i said, its an overall statement, and I made it clear that at Str 7 the lord was worse..though at str 8 or 9 he's better as the Shas'o is instakilled and he is not. At str 10 the problem pops up again where he is actually equal to the shas'o. So i guess its actually only at str. 7(or str 2 i suppose) that the Nec. lord is worse. And even at str 7 he's only arguably worse because of unavoidable special rules.

So for that example I'm quite right, and for other examples my statement is generalist, not absolute, as i mentioned.

 

[Bean]

Look, all I'm saying is that you made this claim: a model's ability to survive fire depends more on its save than on wounds or toughness.

Whether you're making a general statement, an overall statement, or a statement regarding only a set of models, the statement is not true. Whether your tests seem to support it or not (and I'm not saying the math you did is wrong) you've drawn the wrong reason for your conclusions from whatever tests you've done. Models with good armour saves do not tend to be more survivable because the armour save matters more.


Like I said, the armour save does not matter more. I explained why models that otherwise seem comparable (have the same total toughness+successful-save-out-of-six) can have different survivabilities. It isn't because 'the armour save tends to be more important'--it's that the more something deviates towards the extreme ends of the scale for probabilities on the d6, the more impact it has.

Like I said, I don't know what tests you ran, but regardless--the conclusion you've drawn from them and presented here is not true--in any case.