Email in: Erratum [Mathammer on Dreads again]

Monday, November 15, 2010

Email in: Erratum [Mathammer on Dreads again]

Posted by Unknown

"Hey Kirby,

So I feel stupid. My calculations were more than a bit off... (check, check and check again before sending anything :s) Luckily your blog has quite a few readers who are pretty handy with stats.

I feel I owe it to your readers to be more transparant in how I exactly did my calculations so I included an excelsheet. Maybe you clould post it. Then everybody would be free to play around and improve it. Lord knows that there are plenty people out there more adept at Excel to make the sheet user-friendly, for instance.

I would like to ask an additional question:

A friend of mine is playing IG and was wondering which order gives the best avarage result when firing a HWS w/ autocannnons at a vehicle in cover, Fire on my Target (I believe that's reroll succesful cover saves) and Bring it Down (Twin-linked if I am correct). Perhaps your readers could shed some light on this...

Kind regards,

Jabra"

Here's the revised table...

And in regards to your question, mathematically they are the same (assuming 4+ cover) so it doesn't really matter. I'd always do twin-linked (unless 3+ cover) to ensure I'm at least forcing some rolls by my opponent by generating more hits and therefore more likely to glance/pen.

Comments (14)

Logging you in...

Dashboard | Edit profile | Logout

Logged in as

0 Vote up Vote down

wisdom like silence · 756 weeks ago

Not really sure what I'm looking at here. What is "damage"?

2 replies · active 756 weeks ago

0 Vote up Vote down

Jabra · 756 weeks ago

Any roll on the vehicle damage chart. So either a glancing or penetrating hit. I amsumed it would be the result of interest when trying to suppress an enemy army.

0 Vote up Vote down

wisdom like silence · 756 weeks ago

You also aren't using binomial distribution, so your numbers for multi-shot weapons, in terms of the probability of getting at least one useful result, are incorrect.

0 Vote up Vote down

Jabra · 756 weeks ago

Is there any way to post the excel file, Kirby?

1 reply · active less than 1 minute ago

0 Vote up Vote down

Kirby 118p · 756 weeks ago

I couldn't find a way unless I link it via google docs. Anyone else know?

0 Vote up Vote down

wisdom like silence · 756 weeks ago

Here's a spreadsheet that will do this calculation and many others correctly, although I apologise if it's a bit obtuse. I made it for personal use so I tried to explain how it works, but feel free to ask questions about it.
https://docs.google.com/leaf?id=0B9o13REuH5LWNGNl...

It is an excel file (xls).

1 reply · active less than 1 minute ago

0 Vote up Vote down

Jabra · 754 weeks ago

Thanks!

-1 Vote up Vote down

FortyThree · 756 weeks ago

when the odds of the dice you are rerolling are the same (in this case they are, since it's 4+ to hit or 3- to ignore cover) you're always better off applying the reroll the earliest possible, since then you will affect more dice. If you're doing three rolls (to hit, to penetrate, damage), then you'll affect twice as many dice if you use the reroll on to hit than if you use it on to wound. The more unlikely the rolls are the more this is true. This is also the reason why preferred enemy is better than rerolling to wound if the rolls both succeed on the same number

2 replies · active 756 weeks ago

0 Vote up Vote down

wisdom like silence · 756 weeks ago

What do you mean? Statistically the two situations are identical.

+2 Vote up Vote down

Cyklown 51p · 756 weeks ago

Yeah, but the larger the dice pool the more normalized your results will be, even if it doesn't change the over all odds.

+1 Vote up Vote down

FortyThree · 756 weeks ago

umm, nevermind, it seems what I said was completely wrong. Apologies, Kirby is right

1 reply · active 756 weeks ago

0 Vote up Vote down

Kirby 118p · 756 weeks ago

I agree with what you said but in terms of actual numbers over an extended period of time, there is really no difference on the basic scheme of things. I would still encourage individuals to apply the re-roll as early as possible as you indicated though.

0 Vote up Vote down

kaintxu · 756 weeks ago

I'm sorry to tell you that those maths are wrong,

The chance of a TL atuocanon shot is, you need on the die a 3+ and if you fail you reroll for another 3+ which means

0.666666666+0.3333333333333*0.666666666666666 = 0.88888888888888

That's for one of the shots, so for the other one you have the same.

Since both dices (each shot) are independant (you can hit with one not with the other) the numbers go for both to hit as the following

P (A & B) = P(A) * P( B) so 0.88888888888*0.88888888888 = 0.79 to achieve a hit with both sots

The chance to miss a shot is

0.333333333333 * 0.33333333333= 0.11111111111111

This one you got right

The we should calculate either the prob to just hit with one of the shots, or the prob to to hit with at least one of them, which is diferent

1 reply · active 756 weeks ago

0 Vote up Vote down

kaintxu · 756 weeks ago

i dont know how to edit so here is something i left out.

0,11111111111111 is the chance to miss 1 shot, so the cahnce to miss bot shots is

0.11111111111*0.1111111111111 = 0.0122222 = 1.222222%

BTW, all the final numbers on the above post are on a one basis so if you wanna know the % just multiplie by 100