Topic: The problem with Statistics and 40k

[Bannedface666]

What we're saying here is, the REAL game-changers are not the smallest factors,  but the choices you make in game.  The REAL ability to win games at will comes from being able to adapt on your feet with any reasonable list.  Fretting over the little stuff is relatively a waste of time.

Stop telling people this. Some of them might actually understand what you are saying then apply 'fun' into their games.... and I might actually have a challenge!

Last thing I need is people enjoying the game and beating me. I love it when my opponents, only a few of them, go nuts after a loss, wondering how my so and so unit did so well and their elite units failed miserably.

[Draza]

Being a Warhammer site, its inevitable that statistics will be brought up. Statistics can be used to prove points about units/weapons or help bring about new discussions when old one's begin to get boring. Two examples of statistics have been quite enlightening and interesting.

The first being the starcannon vs scatter laser comparison. On paper the starcannon is better at killing marines than a scatter laser. When stats are used, it turns out the scatter laser (for cost) is better at killing Marines than a SC. Interesting to know, when you want to take on Horde/MEQ armies on with the same list.

Secondarily Scorpions and Banshee's have always been compared. Game situations vary, but knowing that Scorpions perform just as well as Banshee's at killing Marines (mainly due to the Exarch) is good to know. Helps with list building, since you know your Scorps can take on those targets.

In the end, statistics basically confirm or disprove our thoughts on units and options. It doesn;t substitute for game experience, but it's ability to help with list making and tactical choices certainly does help (it doesnt hurt after-all) when it comes to being competitive

[Alpha Class Faye]

I love the look on someone's face when they roll four 1's for terminator saves on 4 dice. I love the look on someone's face when they roll 4 1's for that devestator squad of plasma cannons in a building, and they all melt. I love the look on someone's face when the lone guardsman kills the lictor, or the tau fire warriors pummel some genestealers in close combat. 40K's all about those rare moments when the god of all statistics goes on a coffee break, and your creme of the crop unit gets an early shower.

Only last week I had three out of five of my chaos space marine bikers fail difficult terrain tests. Then both were killed by drones in close combat. Happy days

[Ianos Stormbringer]

40K's all about those rare moments when the god of all statistics goes on a coffee break, and your creme of the crop unit gets an early shower.

As you exactly say, its when the Stat God goes for coffee, cause when he is still working you have to work with him to win...

[Sheepz]

Why is ignoring saves good? why is range important? why is the number of shots good?
Becuase it kills stuff.

Way to go. We're almost on the same wavelength.

How do we know how well it kills stuff (so we can tell what is better or more important)?

I feel like we're heading in circles here. You look at the summary sheet. You get the strength, the AP, the number of shots, and the range. Higher the strength = good, longer the range = good, more shots = good, better ap = good.

'But why are they good?'

Because they kill things easier, from further away.

Why? Simply stating one's opinion and simply hoping the other will agree is terribly unconvincing.

Ok, crunch the numbers on a heavy bolter vs a multilaser against Orks and then come back and tell me how minuscule and unimportant the difference is. I didn't need statistics to tell me that, because the guns are practically the same. How one performs is a pretty good indication of how the other will.

Otherwise people would just unreasonably cling to their own opinion without any backup, which would render discussing things useless.

We certainly wouldn't want that, would we?

What is more reliable?

I'm going to stick my neck out and say 'not using them'. At least not using it at the list construction stage, since no doubt someone will make a retarded comment about me doing something stupid in a game on the off chance my rejection of statistics has also lobotomized me and made me a drooling imbecile.

Statistics are generalized. They're not good. You get sketchy ideas from them and that's about as good as it gets. They're boring horrible sums and equations which slow everything down, tax my brain and come to rather unsurprising conclusions (What do you mean it takes 284[lots of] lasguns to kill a Hive Tyrant?!?! What do you mean that the multimelta and the meltagun are THESAME!!!!!)

[Rasmus]

Anyone stating that takes 48* lascannons to kill a landraider is pure wrong, thats not what probablity does.

*(or whatever number of choice)

  Statistics would not claim this, but rather that on average 48000 lascannons will 10000 landraiders.

[Mr.Tanks]

This has sprung into a popular thread. I never knew people felt so passionately about math hammering. It's also nice to see that I'm not the only one who hates it and rather wouldn't use it.

I think however, it is undeniably useful in comparing weapons, and judging outcomes, the use of numbers and probabilities however, is not explicitly required. Many examples have been given on how to compare weapons with out math hammering and as Sheepz said, reviewing the Str, Ap, Range and number of shots gives us a good place to start from.

[Ianos Stormbringer]

Statistics are generalized. They're not good. You get sketchy ideas from them and that's about as good as it gets. They're boring horrible sums and equations which slow everything down, tax my brain and come to rather unsurprising conclusions (What do you mean it takes 284[lots of] lasguns to kill a Hive Tyrant?!?! What do you mean that the multimelta and the meltagun are THESAME!!!!!)

Its not statistics that are generalized, it is your opinion of them. If you rather not use math (which incidentally you do no matter how you put it: you MEASURE range, use a boolean ap, you COMPARE str to toughness or to other str), to the extent other people do, be my guest. True sometimes spending all your time mathing it out instead of playing is the wrong way to go, but so is not doing ANY homework and just play testing through, hunch and superstition. The best IMO is somewhere in the middle.

And sure, there are folks to go overboard and calculate everything statistically disregarding all other factors like metagame, cover, terrain, distances, missions, unit size on the field, etc. Yet that does not mean that knowing a few stats is bad for you especially when there are so many of us mathlovers in the web to provide them for you for free and with lotsa love!

[Flimsyman25]

Ailaros, you are a more patient gentlemen than I.

Folks, it's one thing to say that doing excessive amounts of math detracts from the game because you simply hate math.  That's fine, that's your opinion, and I myself don't care for math.  But to claim that statistics are useless to a game of dice is absurd on the face of it.

Think about it.  You can look at the stats for a weapon.  Look at the range, str., and ap. of a plasma gun vs a bolter.  Same range, better str., better ap.  Plasma gun = better, right?  How do you know it's better?  Exactly as Ailaros said, because it kills stuff better.  How do you know that it kills stuff better?  Because a higher strength (take note here . . .) wounds more often.  Why is the AP better?  Because ignoring armour saves causes more wounds to be inflicted.

You are using statistics, you're just not bothering to do the really simple math.  In many cases, I wouldn't be suprised to discover that many guys who hate statistics so much make lists and play off of the 'common sense' of guys who (for example) swapped Lascannons out for melta-weapons after doing the math . . .

[Benis]

This has sprung into a popular thread. I never knew people felt so passionately about math hammering. It's also nice to see that I'm not the only one who hates it and rather wouldn't use it.

It seems that the hate side is a lot more passionate. People seem to react to statistics as if it is somehow cheating in a debate: If mr A claims that weapon x are better than weapon y against a given target but mr B comes in and says that that is not the case, if you crunch the numbers you will see that generally weapon y is actually better mr A will throw a fit, resort to dismissing statistics, name calling and use arguments as "playing for fun" when it is a discussion about what is best not what is most fun.

I don't know if it is based on an underlying hatred for math or wishful thinking about how good things really are but it is quite annoying to use rational arguments based on probability only to have a comment along the lines of "doesn't count!" thrown in your face.

[Guildmage Aech]

I'd like to draw a subtle distinction between burning our common sense and working out what minuscule advantage one weapon has over a very similar weapon before we ever get to the table.

If you like, but I'm talking about the use of probblity, which frankly IS common sense. Its generally the clueless fools who do dumb stuff all their lives do so because they're unable to make a good estimation of what going to happen as a result of their actions.

Muppets stick forks into electrical sockets do so because they've failed to realised the very high chance of getting electrocuted.

No, i don't think it's considered statistics at all, more of common sense.

Basically my idea is that, when we look at the strength of a weapon, its AP, its range etc, we should already know in our head what its potential is. We know a flamer with its template and AP that it can easily deal with hordes that are closing in range, while a lascannon with its range, AP and strength can in fact take on monstrous creatures. I don't think this is statistics, its' common sense.

Being concious or uncouncious doesn't change what you're doing. You're still using avilable infomation and making a prediction based on that. Its not a difference of black and white, its just stages of grey depending on how much detail you want.

We all know that a multi-laser has pretty much no chance of taking down a holo-falcon in cover, you can look at it and announce' not a chance mate' which doesn't chance the fact you've done math to a degree of accuracty.

A good example is if I threw you a ball to catch, you don't get a rangefinder, work out the balls speed, plot it's trajectory on a graph, calculate the best intercept position to move your hand to, the distance from that position to your hand currently and the orders needed to move your hand... but you do move your hand to where it needs to be and you catch the ball.
You're not psychic and you can't see the future* so the only way you could have known where to put your hand was to have worked out where the ball is going to be.

*Actually you might be psychic, but probablity tells me that its very unlikely...

Because so many things can happen the moment the models hit the table, statistics, and perhaps any pregame planning as such, cannot be relied upon.

No, you can't rely on statistics. As I explained they do not produce reliable predicitions, they produce likely predicitions. Probablity is all about chances, and the thing with chances is that you're always taking a chance.

Critising statistics for not being reliable is like getting rid of your dog because it never does the washing up, such things aren't within it's abilities and are unreasonable to expect from it.

Statistics would not claim this, but rather that on average 48000 lascannons will 10000 landraiders.

Yes, but even so you can't definiatively say they will do it!

[SlaveOne]

I had coached competitive hockey for many years. I used to press upon my players the philosophy, get the puck to a position where the chance of scoring was in a higher percentage and rebounds were likely. While this is fundamental it can be simply countered by the defense guarding the slot and clearing the crease or simply doing there job. Also i began to notice some of my players had the skill of finding the corner of the net from almost any angle. So i changed my philosophy and began to coach based on my players strengths and weaknesses. So instead of using geometry to calculate my percentage in scoring i based it on the player.

Using that analogy, in 40k i play to strength and weaknesses of my units. Playing Eldar for so many years, it was a must.  So i learned that math-hammer will not help me if my banshees are not in position to charge that particular marine sqd. Knowing this, i had to use deployment and movement as these were things i can control. So play-hammer had to prevail. However, i also know that regardless of my knowledge of play-hammer it will not help if i charge those banshees into TH/SS termie sqd of 6 or more, thanks to math-hammer.

[Killing Time]

Knowledge is knowing that a tomato is a fruit.
Wisdom is knowing not to put one into a fruit salad.

Or in other words.
Stats will give you the knowledge about how your units are likely to perform in a given situation.
But only experience will allow you to consistently achieve favourable situations and avoid unfavourable situations.

Dizzy

[eek 107]

What is more reliable?

More appropriate statistical methods... like I said in that very same post.

Actually, if you do standard deviation, rather than simply % chances, you get a range of possibilities that are "normal". While this doesn't tell you the absolute extremes, it does tell you what to expect. In any case, what use would looking at extremes do? On one end, what's the use in telling me that there's the chance that 1,000 lascannons might not kill a single grot, or that 1 guardsman has the long-shot chance of taking out an entire space marine company single handedly?

Fixations with extremes and anomaly isn't all that useful in determining general trends, especially if all things are equally susceptible to the same random factor.

I quite agree that fixation with extremes is inappropriate, which is why I said figuring out what range of results is likely to happen, say 75% of the time, (although I think 90%+ would be more satisfactory).  That still gives you the general trend and is more useful in one specific situation than a single average value.  You can actually expect to acieve the result that you calculate, unlike averages where the chance to achieve the specific result is often pretty slim.  Just how often will 9 Lasgun-armed Imperial Guardsmen rapid-fire exactly 1 Space Marine to death? (rhetorical question, but if you really like doing maths...)
I understand that standard deviation does this, but that's hardly a feature of the conventional mathhammer we see bandied about the internet, is it?
Which is pretty much my point.  The standard practice of mathhammer that we see being used... shouldn't be.  It's place is in your head, during a game, and only there.  When you have the time to work out the statistics, I think it should be done much more thoroughly.

[Ailaros]

Higher the strength = good, longer the range = good, more shots = good, better ap = good.

'But why are they good?'

Because they kill things easier, from further away.

crunch the numbers on a heavy bolter vs a multilaser against Orks

Alright, great. So statistics are the way to go. Good to have this cleared up.

I'm going to stick my neck out and say 'not using them'.

My question wasn't "if statistics are unreliable should we use them?" but "if statistics are unreliable what should you use instead of statistics that is MORE reliable?"

Statistics are generalized. They're not good. You get sketchy ideas from them and that's about as good as it gets.

What gets you more accurate and rigorous ideas?

It seems that the hate side is a lot more passionate. People seem to react to statistics as if it is somehow cheating in a debate: If mr A claims that weapon x are better than weapon y against a given target but mr B comes in and says that that is not the case, if you crunch the numbers you will see that generally weapon y is actually better mr A will throw a fit, resort to dismissing statistics, name calling and use arguments as "playing for fun" when it is a discussion about what is best not what is most fun.

I don't know if it is based on an underlying hatred for math or wishful thinking about how good things really are but it is quite annoying to use rational arguments based on probability only to have a comment along the lines of "doesn't count!" thrown in your face.

Firstly, let me confess that I usually don't like math (my last proper math class was when I was 15, the day math stopped being compulsory), but my general dislike for it doesn't mean that I think it's useless or have a negative attitude towards the things I see as utile (like calculus. Even though I don't know how to use it, it clearly seems to have been valuable over the years).

Other than an irrational dislike of the subject, it probably has more to do with the outcomes. If the results of statistics come up with an idea that competes with your pre-existing ideas (assumedly not based directly on statistics, but some other means), then you're going to blame statistics for coming up with the "wrong" (ie. not yours) idea. Now confirmation bias is everywhere (including this statement (including these two things in double parenthesis)), but that doesn't mean that we can't agree on a system that extends beyond our worldviews to have some sort of way of discussing things without everyone having the same preconceptions. Statistics are a way to do this in 40k.

I think another part of it is, as Dizz and others note, that the actual game of 40k is SO complex with a near infinite number of possible decisions and ways that a game could go that we as humans can't possibly make sense of it all. As such, we rely on such words as intuition, skill, or genius (as Kant put it, the art of doing that for which there are no rules), to make some degree of sense out of a system so complex it can really only be described in vague terms.

That being said, that's not a reason to look at things concretely when we actually can. The process of calculating how far at what angle I should move my troops and what I should shoot them at for the ultimate purpose of winning the game seems subjective, simply because there's no possible way to unweave it all on the table top (or, I might argue, at all, ever). On the other hand, the process of calculating weapon efficiency, etc. is something that is much easier to do. Saying that we shouldn't look at some things specifically just because there is so much we can't look at specifically seems a little odd to me.

Not to say that there isn't a real sense of "intuition" or "knack" or whatever in 40k in general, but it shouldn't exclude us from using real numbers to do real calculations the few times we can.

[blinky jungle(REM)]

2+ armour has twice the chance of succeeding than 3+ armour and that has twice the chance of succeeding than 5+ armour which in turn has twice as much chance than 6+ armour, so 2+ armour has an 8 times more chance of succeeding than 6+ armour. ( I might be talking garbage, but oh well) then you take into account of the AP of weapons than suddenly they become a lot better except the 5 and 6+ armour that in turn gets worse. So thats probabally not very good statistics for all ork players but good for SM with lots of termies.

[Paraplegic]

@ blinky jungle: Uh okay. I'm not sure if anything of that was particularly enlightening towards the discussion but I'm a rather harsh judger of posts.

Stats has been a thorn in my side in any gaming sense. Poker, Diablo 2, any sort of MMORPG will have it's share of statistics to rear it's ugly head at me and suck any enjoyment out of it. That being said, I may not like math hammer, but I certainly won't ignore it when deciding on weapons when building an army. Math hammer does contribute well to some discussions but I"ll admit that there will be times when weapon effectiveness gets brought up and someone posts half a page worth of stats that I'll skip over simply because if I wanted to read the comprehensive statistical analysis of a heavy bolter against a variety of targets, I'll just go play the game instead.

[Impervious]

2+ armour has twice the chance of succeeding than 3+ armour and that has twice the chance of succeeding than 5+ armour which in turn has twice as much chance than 6+ armour, so 2+ armour has an 8 times more chance of succeeding than 6+ armour. ( I might be talking garbage, but oh well) then you take into account of the AP of weapons than suddenly they become a lot better except the 5 and 6+ armour that in turn gets worse. So thats probabally not very good statistics for all ork players but good for SM with lots of termies.

Yes, you're right when you look at it that way, that is one way statistics can appear scewed.

2+ armour will save 5/6 wounds.  6+ armour will save 1/6 wounds.  2+ armour is five times as good as 6+, when you look at it this way. 

[Sheepz]

Its not statistics that are generalized, it is your opinion of them.

Right. Ok. I'm not the only person on this thread to say statistics are good for generalizing.

If you rather not use math (which incidentally you do no matter how you put it: you MEASURE range, use a boolean ap, you COMPARE str to toughness or to other str), to the extent other people do, be my guest

I never said I didn't do maths. I said I don't believe statistics are useful. I also said I believe they shouldn't be used in the army list competition stage unless you're some stupidly competitive Sirwinalot. Because they're largely unhelpful. Furthermore, they bog down threads and are used to needless ends.

Folks, it's one thing to say that doing excessive amounts of math detracts from the game because you simply hate math.  That's fine, that's your opinion, and I myself don't care for math.  But to claim that statistics are useless to a game of dice is absurd on the face of it.

1) I don't hate math.
2) I did not claim they were useless. I claimed they were relative to the desired outcome and failed to take into account a dozen mitigating factors.

Alright, great. So statistics are the way to go. Good to have this cleared up.

Man, I know we disagree, but there's no reason to butcher a post and take it out of context to try and win an argument. Run the statistics on a heavy bolter vs a multilaser. What are the numbers you get. Miniscule and frankly unimportant in the grand scheme, because I can never roll .48 of a dice.

My question wasn't "if statistics are unreliable should we use them?" but "if statistics are unreliable what should you use instead of statistics that is MORE reliable?"

And my question is. Why must there be some sort of scientific method for approaching a game? Why must it be taken so seriously and why, oh god why, are we bothering to calculate such precise and useless numbers?

What gets you more accurate and rigorous ideas?

Sticking my head down a toilet filled with genetically modified, venomous, piranha's. What on earth do you mean by 'rigiourous ideas'?
 

It seems that the hate side is a lot more passionate. People seem to react to statistics as if it is somehow cheating in a debate: If mr A claims that weapon x are better than weapon y against a given target but mr B comes in and says that that is not the case, if you crunch the numbers you will see that generally weapon y is actually better mr A will throw a fit, resort to dismissing statistics, name calling and use arguments as "playing for fun" when it is a discussion about what is best not what is most fun.

I don't know if it is based on an underlying hatred for math or wishful thinking about how good things really are but it is quite annoying to use rational arguments based on probability only to have a comment along the lines of "doesn't count!" thrown in your face.

What the hell is with all these psychiatric evaluations? If you use maths to prove a weapon is better than another, good for you. I'm not really going to kick up a fuss. But when you use maths to weigh up niggling details that aren't important, I'm going to get irritated. Especially when you have to use increasingly bizzare situational evidence to support the maths. You start with A heavy bolter shooting at a T3, 5+ save unit. Then you go, "Okay, what if they're in cover" and do it again. Then you do it against a 3+ save unit. Then you do it against a 3+ save unit with FNP. Those two are badly constructed comments, but stuff like, "Okay, I'm running the numbers on Ogryns vs Marines." Then you try and work out one side charging, then the other, factor in shots before charging, factor in powerfists, factor in variable squad sizes, factor in cover and frag grenades and all that the Udders of Thoth until you've finally conclusively proved that Ogryns don't do well against 25 Space Marines with Powerfists and Feel No Pain, led by a Chaplain (and they're charging). Brilliant stuff. Just fantastic. And needless. Because it doesn't matter. These hypothetical situations one uses to prove or disprove the awesomeness of a unit never really comes of just like that. And when they do, probability dictates that it's not always going to be 'average'.

Stats has been a thorn in my side in any gaming sense. Poker, Diablo 2, any sort of MMORPG will have it's share of statistics to rear it's ugly head at me and suck any enjoyment out of it. That being said, I may not like math hammer, but I certainly won't ignore it when deciding on weapons when building an army. Math hammer does contribute well to some discussions but I"ll admit that there will be times when weapon effectiveness gets brought up and someone posts half a page worth of stats that I'll skip over simply because if I wanted to read the comprehensive statistical analysis of a heavy bolter against a variety of targets, I'll just go play the game instead.

Which summarizes to a certain degree my view of mathhammer, before people accuse me of being ANTIMATH! or 'hating math', or just mention that I'm totally wrong and resorting to childish insults because I didn't have sex with my mother or something...

[NewHeretic]

And my question is. Why must there be some sort of scientific method for approaching a game? Why must it be taken so seriously and why, oh god why, are we bothering to calculate such precise and useless numbers?

If I may answer your questions, Sheepz...

There needn't be anything of the sort.  We (meaning you) don't take statistics so seriously in this game -- only some of us do. 

You see, such statistical analyses are engaged in to different degrees by those of us who actually enjoy utilizing them in our attempts to be as competitive as possible.  Yes, it seems to me that the more competitive a player is, the more likely s/he is to take advantage of anything that will provide him/her an edge in the game (including the issue at hand).  This is a generalization, of course, and doesn't apply universally, but, as a generalization, I believe it is true.

Really, if you don't enjoy involving statistical analysis in your approach to 40k, then don't.  Why worry if your opponent or others on an Internet forum seem to be obsessed with it?  You are free to enjoy the game as you wish, and so are we they. 

Seems simple, huh?

NewHeretic