G.K. Librarian and Dreadknight

[Baka]

Casting Might on the Dreadknight would bring his Str to 7 with one D6 added to armor penetration? So does that give him make it 7+2D6?

[Diagnosis Ninja]

Casting Might on the Dreadknight would bring his Str to 7 with one D6 added to armor penetration? So does that give him make it 7+2D6?
Check out the Monstrous Creature rules in the main rulebook, and you'll be plesantly surprised to find it's a little bit different to that. Use it with the Greatsword ;)

[memnarch_129]

Diagnosis is correct it is much more than the 7+2D6. If you use the Dreadknights Hammerhand in addition to the Librarians Might of Titan the Dreadknight is actually rolling 8+3D6 (2D6 for being a monstrous creature and one from MoT). As Diagnosis point out try it with the Greatsword, rerolling ALL failed to hit to wound and Armour Penetration rolls is somthing your opponent hates to see.

[Demonus]

Still dies to 1 raider full of splinter rifles =)

[Archon Charybdis]

Still dies to 1 raider full of splinter rifles =)

You must have some *****in' *** raiders to fit 72 splinter shots into one raider.

[dannyat2460]

You must have some *****in' *** raiders to fit 72 splinter shots into one raider.

Or luck ive seen it taken down in combat by guardsmen but that was a complete fluke

[justsam]

i once saw a daemon prince die to shotguns. at that point, i stopped categorizing any death as reasonable or unreasonable.

[Demonus]

Oh Ive seen all sorts of things die to damned 45 lasgun shots.

As far as splinter rifles go, they are 4+ wound vs Dreadknights right? A squad of 10 double firing has a pretty good chance of killing a dreadknight. Especially when the GK player is prone to rolling 1s =P

[DarkLink]

No, it's not good at killing Dreadknights. Sure, you ignore the T6. But they still have 2+ armor saves.

There's way too much math involved to figure out what exactly the odds are, but you can't just assume that the GK player is going to roll a bunch of ones.

If you look at the averages, it takes four wounds to kill, which requires 24 armor saves, which requires 48 wounds, which requires 72 shots as mentioned above. 50% of the time, it takes 72 or more splinter rifle shots to kill a dreadknight. A raider full of warriors gets 20. Plus a dark lance or two I guess, but the point is, the odds of a single raider killing a dreadknight with splinter rifle fire is pretty poor (quite a bit less than 50%, though we don't know exactly what it is).

Common sense might tell you that a 4+ to wound would be good against Dreadknights since it ignores T6, but common sense does not always line up with reality. Common sense can give you a hypothesis, but you need something to back that hypothesis up. In this case math is sufficient, and the math says the odds are not very good.

[Demonus]

Soo what you are saying is, bring a second raider full of warriors with splinter rifles...gotcha! :D

[Nungunz]

There's way too much math involved to figure out what exactly the odds are, but you can't just assume that the GK player is going to roll a bunch of ones.

(2/3) * (1/2) * (1/6) = 5.56% chance of causing a wound with a splinter rifle shot.


So as you say, an average of 72 shots to kill a dreadknight. Averages generally don't tend to mean jack on an individual basis, though. Gotta figure in standard deviations, and all the jazz.


Soo what you are saying is, bring a second raider full of warriors with splinter rifles...gotcha! :D

Considering that a dark lance and 20 splinter shots does an average of 1.48 wounds to a dread knight, I'd bring a third as well.

;)

[Diagnosis Ninja]

If you look at the averages, it takes four wounds to kill, which requires 24 armor saves, which requires 48 wounds, which requires 72 shots as mentioned above. 50% of the time, it takes 72 or more splinter rifle shots to kill a dreadknight. A raider full of warriors gets 20. Plus a dark lance or two I guess, but the point is, the odds of a single raider killing a dreadknight with splinter rifle fire is pretty poor (quite a bit less than 50%, though we don't know exactly what it is).

So? I might just have the numbers entirely wrong, but if you factor in:

BS 8
4+ to wound
AP 1 and a 4+ cover save

Doesn't a Vindicare Assassin only wound 40% of the time against anything?

The point is, that you can't expect mathematics to stand up in every game. Sure, if you look across the entire planet, it wouldn't surprise me if that was the average statistic, but the entire thing is governed by luck. I might never wound a Dreadknight with a Raider full of Blasters and Splinter Weapons, but I might kill it.

On top of that, you're looking at a straight vacuum. There's an entire army also there. If a Dark Eldar player really wanted a Dreadknight to turn into a Deadknight, they could manage it easily. Of course, that would be at the peril of ignoring the other Grey Knights present.

You aren't dealing with a predictable thing. You have no idea how the other player will respond to circumstances, even if you set things up as best you can. They might still just decide to do it anyway.

Sure, use maths as a guideline, but don't expect it to hold true every time.

[Demonus]

Yep Im not a believe of Mathhammer at all. I watched a player roll 6 dice last tourney. he rolled 6 1s. I faced another player with 5 chosen and a rhino with a combi melta. Thats 6 melta gun shots against his soul grinder. I managed 1 pen, and all it did was take off a weapon.

The next few rounds, I laughed as his Soul Grinder took 3 turns to kill 5 chosen, with 0 chance to do anything to him, and no save should he wound me.

Too much randomness to take into account in 40k.

[DarkLink]

(2/3) * (1/2) * (1/6) = 5.56% chance of causing a wound with a splinter rifle shot.

You haven't take a statistics class, have you.

If a single splinter rifle shoots at the dreadknight, then yes, you're correct, there is a 5.56% chance of causing a wound.

However, if you take twenty splinter shots then the odds are much, much, much more complicated to calculate. Here's the wiki page (http://en.wikipedia.org/wiki/Binomial_probability) for the long explanation, I'll explain it in brief.

To calculate odds for a large number of trials, the formula follows this form: (n choose k)*p^k*q^(n-k). n is the number of trials, k is the number of successes, p is the probability of success and q is the probability of failure. Now, if you only have a single trial (n=1), then the formula simplifies down to the form you use above. This formula requires that you define k, the number of success. So you can calculate the probability of getting exactly 16 hits out of 20 trials at those probabilities.


However, if you want to move on past that, it gets really cumbersome. You have to calculate the odds of getting 0 hits, 1 hit, 2 hits, 3 hits... all the way up to 20 hits. That's 21 uses of the formula. Now you have the odds of getting any particular number of hits.

Now that you have the number of hits, you have to reuse the formula to get the odds for each number of wounds. 0 hits causes 0 wounds. 1 hit means one trail means 5.56 odds of causing a wound. 2 hits means you need to use that formula two more times, in order to get the odds of causing 0, 1 or 2 wounds given two hits. 3 hits means 3 more calculations to find the odds of 0, 1, 2 or 3 wounds given 3 hits. And so on until youve gotten all the way up to calculating the odds of 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 wounds given 20 hits.

Now that you've done that, you need to multiply through and sum stuff up so that you now have the odds of getting 0, 1, 2, 3, 4... up to 20 wounds.


Now, you have to repeat that whole process, but for how many unsaved wounds given the number of wounds caused, then multiply though and sum up everything again until you have the probability of causing 0, 1, 2, 3 or 4 wounds.



Even just figuring out how many iterations of that formula is a chore. So take it from me, since I have taken statistics in college. It is not an easy calculation to make.



As for the standard deviation, it's not quite that simple either. For one thing, I don't even know what the formula for the standard deviation of 20 d6 dice is, and even if I did know you would still have to do some tricky math to translate that through from the odds to hit to the odds to wound to the odds of failed saves.



So, again, not this isn't easy or simple math. Well, ok, it's simple math, you just have to do so much bookkeeping that it isn't easy. You feel free to do it if you want.




So as you say, an average of 72 shots to kill a dreadknight. Averages generally don't tend to mean jack on an individual basis, though. Gotta figure in standard deviations, and all the jazz.


Averages actually mean quite a bit. By definition, without looking at standard deviation or anything, 50% of the time it takes 72 splinter shots to kill a Dreadknight. That inherently tells me the odds are not that good, even if there is a really big standard deviation. If we assume a standard distribution (which may or may not be valid for a large sample of dice rolls, I don't know), then 20 shots is roughly somewhere around 2 standard deviations away from the mean. That means that there is roughly a 15% chance that 20 or fewer shots will successfully kill a Dreadnought. Now, keep in mind that the math is pretty rough and the assumptions leading to it might not be valid, but it's probably a fairly accurate ballpark figure. It might be 10%, it might be 20%, but the point is the odds aren't good.



Considering that a dark lance and 20 splinter shots does an average of 1.48 wounds to a dread knight, I'd bring a third as well.


If you want to go by averages, then we can do that.

Let's say 10 warriors with 1 dark lance/blaster, in a raider with a dark lance. That's 2 dark lances plus 9 splinter rifles.
2*(2/3)*(5/6)*(2/3)=.74 wounds
9*2*(2/3)*(1/2)*(1/6)=1 wound

That's about 160pts within 12" range causing 1.74 wounds.



Alternatively, let's look at a ravager with 3 dark lances.
3*(2/3)*(5/6)*(2/3)=1.11 wounds.

Now, that's only, what, 115pts? So we'll normalize that to the 160pts, which brings us up to 1.55 wounds. So, it seems that 9 splinter rifles is better than 1 extra dark lance, but not by a whole lot. Plus the ravager can put wounds on from 36" as opposed to 12".

Long story short, neither are what I'd call ideal for shooting dreadknights to death, but both can do the trick. Ravagers have the advantage of range, mobility and extra anti-vehicle, while warriors have more wounds and more anti-infantry.

But my point is, if you want to shoot the dreadknight to death you should probably go with triple disintegrator ravagers or something like that. Maybe trueborn with four blasters, backed up by some potshots from elsewhere.


So? I might just have the numbers entirely wrong, but if you factor in:

BS 8
4+ to wound
AP 1 and a 4+ cover save

Doesn't a Vindicare Assassin only wound 40% of the time against anything?

Depends on what ammo you're using, the odds change. He has the options of a 2+ to wound (.76% chance of wounding), removing invulnerable saves (.917% chance), and 4d6 vehicle penetration. With all the argument over stuff like whether it's 3+4D6 or just straight 4d6, that can cause issues.

But long story short, the vindicare is awesome against 1 wound infantry or dudes with invulnerable saves, but yes he only has like a 30% chance of killing a land raider.



The point is, that you can't expect mathematics to stand up in every game.

Yes. Yes you can. There's a very large and supremely important field of mathematics upon which most of modern science is based that inherently proves that you are wrong about this. A massive portion of our knowledge about science, engineering, physics, biology, disease and medicine, marketing, politics, manufacturing and just about everything else is based on statistics. If statistics didn't work, then none of that other stuff would work either. There is such an overwhelming body of evidence that statistics works that any anecdotal evidence you could try and muster up will be entirely irrelevant. Just because you think that you always seem to roll 1's at the wrong time does not mean that statistics somehow magically does not apply to you.


Statistics works. You just misunderstand what it tells you. It doesn't tell you that it always takes 72 splinter rifle shots to kill a dreadknight. It means that on average it takes 72 splinter rifle shots to kill a dreadknight. That is always, always, always true, unless you are using loaded dice, and if you don't understand the difference between those two statements then you need to go do some studying on how statistics works.




Wargamers always seem so utterly ignorant of the foundational mathematics behind statistics that it always tempts me to write an article up on basic statistics as it applies to wargaming. But I'm lazy, so I just tell people that they're wrong instead :p.

[Diagnosis Ninja]

But that's my point, it's an average. It won't happen in one game, but the pattern will be there in the long run. But, you seem to think that I'm entirely ignorant of, well, anything and everything seen as you also started to pull physics and all that into it, so I'll just let you get on with it.

[DarkLink]

Actually, it was more about correcting the math errors in Nungunz's assumption and Demonus' faulty idea that mathhammer is somehow invalid than what you were saying. As you say, mathhammer is just an average, a long term thing rather than an isolated event. Rolling 6 1's is well within the reasonable range for mathhammer, because it doesn't happen very often.


My point, however, is exactly that. It doesn't happen very often. If I bet you money right now that you couldn't pick up six dice and roll all 1's, would you take my bet? It's possible that it could happen, sure. But only an idiot would bet on it.

Same thing with shooting at the dreadknight. Betting on 10 warrior killing a Dreadknight is a bad idea, because the odds are not in your favor. If you asked me to bet on that, I would quickly do the math and realize that the odds are in favor of the Dreadknight surviving and put my money there.

That's what mathhammer tells you, and that's why I wouldn't rely on 10 warriors shooting a dreadknight to death.

[Tynskel]

ack, talking about long term averages! You are starting to sound like you are talking about the differences between climate and weather, and how climate is predictable due to long term averages!

[CitizenSoldier]

At 5.56% probability to wound, you have 2.26% probability of taking 4 or more wounds with 20 splinter rifles.

(over many attempts about) 1 in 44 people who try it will succeed. The people who witness 43 of these attempts will quickly forget such a non-event. The people who witness the success will likely not forget it so quickly, leading to the impression that it happens far more often that it actually does. Our brains work like this - they remember the anomalies in life.

One is fact, the other is perception - you're free to follow whichever you want.

Citizen.

[Tynskel]

Ack, now you sound like someone who saw a storm and is calling it climate change (or the opposite)!

[wolflold]

Dreadknights are cool but ill never take one again! Played a few games against Dark Eldar, T6 isn't going to do any good against Agonisers! Killed the dreadknight everytime in a single round :(.

[Demonus]

Doubles game. GK vs Chaos Marines. Had DK survive 2 Melta Gun shots, CC with a Defiler, and then 6 Plasma Gun shots. Finally succomed to 10 berserker charge.

Partner was rolling insane for his inv saves.