Forum adverts like this one are shown to any user who is not logged in. Join us by filling out a tiny 3 field form and you will get your own, free, dakka user account which gives a good range of benefits to you: No adverts like this in the forums anymore. Times and dates in your local timezone. Full tracking of what you have read so you can skip to your first unread post, easily see what has changed since you last logged in, and easily see what is new at a glance. Email notifications for threads you want to watch closely. Being a part of the oldest wargaming community on the net. If you are already a member then feel free to login now.

Yeah, but studies show that only 49% of Mark Twain quotes are true...

at least there's that

IIRC Twain was quoting some British Lord or PM...

Again, I have no problem if you run the numbers in your head, but pulling out a calculator during games does two things. First, it ruins any modicum of fantasy either player may have and second, it does tend to pull away from the tactical way to play. Besides, there are WAY too many variables to put into a calculator, so what's the point?

To the poster who said that real generals don't do math on the battlefield, I say, you're right.

They have teams of ballistics experts running numbers in an HQ somewhere, who pipe that information to the general to inform his decisions. We 40K players just have to do it for ourselves (and it is FAR less complicated than real-world calculations).

I refer you to this article:
http://query.nytimes.com/mem/archive-free/pdf?_r=1&res=9C04E2D8113AE63BBC4851DFB3668383669FDE

This is just the first result I found of MANY that shows the kind of abstract statistics and probability that are applied every single day of warfare, and have been since the days of the Persian Janissaries (calculating quantities of gunpowder to length of fuse, etc.). I've seen wall-size posters with pictures of ordnance size to plating thickness that can be *reasonably expected* to be penetrated given a direct hit after accounting for armor slope.

If you don't respect or understand the use of probability and statistics in warfare (or wargaming, where it applies even more directly), then you obviously have no idea what it takes to wage modern war. Modern generals do not guess. They do not fling ammunition at targets to "see if it works". They calculate what should do the job, then up it a little bit to be absolutely sure. Even then, things go wrong, but more often than not, doing the homework pays off bigtime.

In 40K, Mathhammer is just a fancy term for the statistical analysis of the result of average rolls. Everyone who uses it should know that dice are random, so there is a huge range of possiblity, but knowing the *probability* is how you determine your application of force density.

If you overkill a unit by positioning too many of your units to kill it, but it only takes half, you've wasted the shooting AND positioning of the other half of your units. Knowing roughly what a unit will do in a given situation allows you to prevent wasting force, or from relying on a long-shot by under-committing. Really, like everyone says, it *informs* your decision, it does not make it for you.

I begin every game with a simple prayer to the Dice Gods.

"Please, oh please, please oh Gods, just let me roll average."

*applaud*

Thank you! Also, I was the one who posted about general and math in the field. I'm well aware that people do do pre-war calculations, and that what I was arguing for. I'm against in-battle calcs.
All praise to the Dice Gods!

I stand by my previous statement, mathhammer will only frustrate you let me re-phrase that it will frustrate me! If I play another game and some one says the dice screwed them because on average I am going to scream.

I just like mathhammer to know how much I can complain about dice rolls. When I get 30 attacks from charging CSM and kill zero enemies, I know I did pretty darn badly. Or when those same charging marines lose combat to fire dragons in hth.

It's not unreasonable to say "27 bolter shots should kill about 3 marines". It really isn't. So when I'm moving in the movement phase and I want a squad dead, I'll position myself to put the right amount of firepower on them. What I'll also do is have a couple backup units that can shoot them if I roll badly, but also have other targets if things go as planned. Rarely does it work out perfectly. There'll be that last marine with a meltagun that you KNOW is going to blow up your landraider if you don't kill him. But having fired all your firepower you've allocated, he's still alive. Now you use your reserve and....still alive. Now you turn your long-range lascannon devastator squad you were going to use to take out that dakka predator on him because you KNOW he's got your landraider in his sights and he still won't die!

These things happen.

My goal in the Movement phase is always to set up a multi-layered shooting plan. Priority targets will be surrounded so they are in the kill-zone of multiple units in case the "normal" amount of force doesn't finish the job. If it does, they have been placed so that they can direct their fire elsewhere. This tends to lead me into a staggered advance, keeping units inside one another's threat radius, and thereby protection. Balancing these offensive maneuvers with positioning designed to protect me from the upcoming enemy phases is the greatest tactical challenge in this game, IMHO.

If we consider that probability means a single grot can take on a fully kitted terminator unit, it makes it apparent how silly the "probability negates your mathammer" argument is. Sure, the grots /can/ kill the terminators. However I'm not going to let the game hinge on that. I might, however, let the game hinge on three Vindicator pieplates dropping on the Terminators' heads.

jimbob123432 wrote:Thank you! Also, I was the one who posted about general and math in the field. I'm well aware that people do do pre-war calculations, and that what I was arguing for. I'm against in-battle calcs.
All praise to the Dice Gods!

You keep bringing up this wierd point. I'm wondering, how many games have you played against people who pull out calculators during the game? I haven't played, met or ever seen any... and I doubt many (if any) people here have.

I have never PLAYED against someone using a calculator but I have watched games where it happens.

The only problem with mathhammer is that it is like a science experiment under perfect conditions the results do not reflect in game results.

Translation. "I don't really understand X, so it must be stupid."

It's already been beaten to death what Mathhammer does for you. If you don't get it, you don't get it, but that doesn't make it any less useful to people who do. So let's try to get a more actionable take away from this thread.

Here's what I'd suggest: Don't negatively judge something you don't understand.

Or: Don't negatively judge something that is entirely optional, and nobody is forcing you to do.

That, right there, is two very good reasons for your post to never have existed.

The best reason, though, is probably that it doesn't just make you look confused, it makes it apparent that you're frustrated enough with your confusion that you needed to post about it.

but experience is better than going by bad math... and thats what math hammer generally is.

It's not "bad math." There are so many false assumptions here...

First off, you seem to think that "Mathhammer" means basic multiplication of odds and trials to provide an average outcome. It doesn't have to mean just that. It can mean any form of mathematical examination of probabilities and statistics in 40K.

Regardless, even if we're talking about the simple approximations generated by muliplication of odds and trials, these approximations are still fairly accurate. They're not as comprehensive as a full binomial distribution, but they nicely capture the expectation of the situation. They're far more accurate than the "feeling" one might amass over any number of games.

If nothing else, they're a common language. If you tell somebody that a weapon kills "about 3 MEqs" they know a lot more clearly what you mean than if you said "a good number of MEqs."

In real combat, generals don't pull out calculators and determine the probability of their snipers taking out an infantry unit.

Actually, they do.

Well, not exactly, but being able to reduce things to expected outcomes is very much a part of military leadership. They run simulations, determine what their expectations are, etc. While they don't, say, calculate the average bench press of a given infantry squad, and divide that by the number of machine guns, they do try to determine the expected outcomes of given situations.

But, whatever. Even if they didn't, this isn't real combat, it's a game based on... Ohhh, that's right. Probabilities and statistics.

Let's instead compare it to something that's far more analogous: Casinos.

40K is a whole lot more like a casino game than it is like real warfare. I hope it's obvious just how much effort casino operators and game players put into examining those probabilities and statistics.

I fully agree with Phryix

LunaHound wrote:I fully agree with Phryix

+1

imweasel wrote:

LunaHound wrote:I fully agree with Phryix

+1

More like 1/3*1/3+1

jimbob123432 wrote:Again, I have no problem if you run the numbers in your head, but pulling out a calculator during games does two things. First, it ruins any modicum of fantasy either player may have and second, it does tend to pull away from the tactical way to play. Besides, there are WAY too many variables to put into a calculator, so what's the point?

I agree, if your gonna pull out a calculator in the middle of the game, and start evaluating your chances, rather than just do it in your head. then id be a bit put out.
I always fallow the good rule of 50/50. you do or you dont, of course playing and knowing what your army is capable of will win for you a hell of alot more than some pencil necked geek whipping out the calc in the middle of the game.

you have no idea what you're talking about. admit it.

Obvious troll is obvious.

Phryxis +1, 'atta boy!

+1 Phryxis.

I'll be honest. I make my lists in excel, usually have my laptop with me to make changes at the store (copying to paper for the actual list.)

One tab of excel has a basic template for a binomial distribution built in. Have I ever used it in a game? Well, actually once, but only because my opponent asked me to. It was more of a "what are the chances" thing, and he knew I had the program and wanted to know.

I set up my lists based on probabilities because I don't give two turds what the average is, I want to know what is "likely." I base my probabilities on importance. If I NEED something to work, I'l pull it out to 90% If I would really like something to work, I'll just pull it out to one standard deviation.

For instance, I NEED to stop a Rhino with my (Tau) Missile Pods. So I plug the numbers into my spreadsheet, and I find out that three battlesuits with TLMP and TA have a little over 67% of rolling an immobilized result on the vehicle damage chart. To me, that is not good enough. If I bring in two Broadsides as back-up, they have roughly the same chance as the three deathrains, giving me a combined probability that is much more closer to my acceptable limit, plus they have a chance of taking out that land-raider after the Rhino is popped, allowing my Deathrains to rain... death on the Marines that popped out of their Rhino 12" off my opponents board edge.

Obviously that is just the tip of the iceberg, but even though I don't make true probability calculations in my head (with the exception of ones I have memorized) it really helps me allocate my forces effectively pre-game.

One thing that I absolutely despise (and other people have mentioned) is the passing of statistical means as probabilities. People, they are NOT probabilities and should not be used as such unless you are either rolling a crap-ton of dice or understand that statistical averages are almost always over-estimates of the probable outcomes. My most recent example is the Tervigon. I started a nid army, so I wanted to see what this monster was made of. Statistically is should poop out 23 Termagants before it goes kaput, which is nice, but that is so horribly skewed by the large results from unlikely rolls further along the distribution that it can be misleading. The P(max) is more like 14 and you can not reliably count on more than 16-18 per game. The reason the average is so high is because on the off-chance that you do get 3,4 or 5 rolls in a game the numbers are HUGE, pulling the mean far far far to the right.

Also, +1 Phryxis

I'm probably using the wrong terms here, because it's been several years since I've tried to do any statistics, but:

Something I've always wanted to see but haven't is a discussion of the standard deviation of various weapons or units. For example, according to arithmetic mathhammer, it should take about 30 bolter hits (45 shots at BS4) to kill 5 marines. It should take 6 plasma hits (9 shots at BS4, 12 at BS3) or 18 autocannon hits (36 at BS3) to kill the same 5.

My question is, which weapon has the lowest standard deviation? In other words, which is most likely to kill exactly or very close to the 5 marines?

This would be very helpful, I think, because it would let you a) choose weapons that fit your style of play (risky or dependable), b) better allocate shooting during the game (i.e. put 'dependable' weapons on the really important targets, shoot the 'risky' ones at targets of opportunity).

Also, Fizyx, that program sounds very helpful. would you mind posting or PMing the details?

+1 Phryxis

Also, I find Mathhammer to be extremely helpful in deciding what I want to field. Several of us have gone so far as to make Binomial Distribution programs which graph the most likely outcomes of a given situation.

Frankly we all think its kinda fun.

Statistically is should poop out 23 Termagants before it goes kaput, which is nice, but that is so horribly skewed by the large results from unlikely rolls further along the distribution that it can be misleading. The P(max) is more like 14 and you can not reliably count on more than 16-18 per game.

I'd like a little more detail on this... I don't see how anything is getting skewed. Not that you're wrong, just that I don't understand what you mean. To me, the 23-24 figure seems perfectly reasonable.

My question is, which weapon has the lowest standard deviation? In other words, which is most likely to kill exactly or very close to the 5 marines?

Looking at the binomial distribution helps a lot with this. Typically there are a cluster of results that are likely, and then nothing outside of that. You're almost always dealing with a bell curve, and then your question is basically asking how flat the curve is.

I think it's important to be able to handle all the possible results: it may not be likely to wipe out a Terminator squad with a single round of Lasgun fire, but you have to be able to handle it all the same.

For example:

Which would you prefer? A Sx APy weapon that is Heavy 2, or a Sx APy weapon that is Heavy 1 Twin-Linked?

You should prefer the former, because while the latter has the same likelihood of causing a single hit, it has a greater potential for damage.

Fizyx wrote:Statistically is should poop out 23 Termagants before it goes kaput, which is nice, but that is so horribly skewed by the large results from unlikely rolls further along the distribution that it can be misleading. The P(max) is more like 14 and you can not reliably count on more than 16-18 per game. The reason the average is so high is because on the off-chance that you do get 3,4 or 5 rolls in a game the numbers are HUGE, pulling the mean far far far to the right.

I disagree with you. Probability would say that you should get more than one successful spawn. Two spawns (the smallest number that is more than one) should yield more than 16-18 total Gants simply because of the 7+x distribution on 3d6 is going to be more common. 19-21 isn't much more than 16-18, but it certainly is more.

Now, I do agree with you that the incredibly small number of rolls you can expect to make with a Tervigon before it burns out (more than one but less than three) means you really can't bank on the bell curve being normalized. This doesn't push the mean in any direction, though (maybe to the left since you can burn out and therefore your results will be skewed lower by no longer having opportunity to normalize the curve).

So... to be good at warhammer, you have to be good at math? What about people who do absolutely no math during the game, and are terrible at math in general?

BobTheChainsaw wrote:So... to be good at warhammer, you have to be good at math? What about people who do absolutely no math during the game, and are terrible at math in general?

Some players instinctively realize what they have to do and what to do it with.

If you don't have that, and are terrible at math, there is a common term for those players.

Food.

sourclams wrote:

Fizyx wrote:Statistically is should poop out 23 Termagants before it goes kaput, which is nice, but that is so horribly skewed by the large results from unlikely rolls further along the distribution that it can be misleading. The P(max) is more like 14 and you can not reliably count on more than 16-18 per game. The reason the average is so high is because on the off-chance that you do get 3,4 or 5 rolls in a game the numbers are HUGE, pulling the mean far far far to the right.

I disagree with you. Probability would say that you should get more than one successful spawn. Two spawns (the smallest number that is more than one) should yield more than 16-18 total Gants simply because of the 7+x distribution on 3d6 is going to be more common. 19-21 isn't much more than 16-18, but it certainly is more.

Now, I do agree with you that the incredibly small number of rolls you can expect to make with a Tervigon before it burns out (more than one but less than three) means you really can't bank on the bell curve being normalized. This doesn't push the mean in any direction, though (maybe to the left since you can burn out and therefore your results will be skewed lower by no longer having opportunity to normalize the curve).

I ran a numerical analysis and plotted a histogram. There is a sharp drop-off right at the 16-18 mark. The graph is on BOLS forums.

Fizyx:

Links please.

http://www.lounge.belloflostsouls.net/showthread.php?t=5295