Switch Theme:

Probability vs Statistics  [RSS] Share on facebook Share on Twitter Submit to Reddit
»
Author Message
Advert


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.




Made in se
Been Around the Block





Probability or statistics when you build your lists? Never been one to have much success going by probability and what I could potentially achieve as others often suggests to me. Statistically I get all the 1s when rolling to hit and my opponent get all the 6s when rolling to save, add all our rolls together and math for probability checks out on the whole. Solution, higher AP weapon to not allow the opponent to roll for save when I do get hits through. Works decently untill i meet an opponent with lots of inv and other counters. Trying to counter it wiht lots of mortal wounds but so fore its been unreliable for the most part. Thinking of trying out some Ld debuffs aswell to get more out of my successes. Rerolls are not a safe bet as i have had sets of 1s rerolled with a new set of dice shaking it up properly and ending upp more 1s. Though I have been told that I am rather obnoxious to play against no matter what list or army I use. Presumebly becuse I am "better at the game" witch i have a hard time interpreting. So other than rubbing my opponent of their Sv rolls, forcing morale fail and mortal wound bomb them what can I do to eliminate my opponents units reliably when probability keeps failing you? Are there better ways to mitigate probability?

Deathwatch +3000p
Farsight +2000p
Kraken +2000p
Nephrekh +1000 
   
Made in us
Decrepit Dakkanaut




Statistics that rely on anecdotes are merely cute and that's it. Just because your Berserker Marines get 6's on all their roles when you shoot with then doesn't mean they're good at shooting.

CaptainStabby wrote:
If Tyberos falls and needs to catch himself it's because the ground needed killing.

 jy2 wrote:
BTW, I can't wait to run Double-D-thirsters! Man, just thinking about it gets me Khorney.

 vipoid wrote:
Indeed - what sort of bastard would want to use their codex?

 MarsNZ wrote:
ITT: SoB players upset that they're receiving the same condescending treatment that they've doled out in every CSM thread ever.
 
   
Made in se
Been Around the Block





Slayer-Fan123 wrote:
Statistics that rely on anecdotes are merely cute and that's it. Just because your Berserker Marines get 6's on all their roles when you shoot with then doesn't mean they're good at shooting.
It has held true and proven to be useful for my list building.

This message was edited 2 times. Last update was at 2020/04/02 01:54:51


Deathwatch +3000p
Farsight +2000p
Kraken +2000p
Nephrekh +1000 
   
Made in us
Unrelenting Rubric Terminator of Tzeentch





Gaen wrote:
Probability or statistics when you build your lists? Never been one to have much success going by probability and what I could potentially achieve as others often suggests to me. Statistically I get all the 1s when rolling to hit and my opponent get all the 6s when rolling to save, add all our rolls together and math for probability checks out on the whole. Solution, higher AP weapon to not allow the opponent to roll for save when I do get hits through. Works decently untill i meet an opponent with lots of inv and other counters. Trying to counter it wiht lots of mortal wounds but so fore its been unreliable for the most part. Thinking of trying out some Ld debuffs aswell to get more out of my successes. Rerolls are not a safe bet as i have had sets of 1s rerolled with a new set of dice shaking it up properly and ending upp more 1s. Though I have been told that I am rather obnoxious to play against no matter what list or army I use. Presumebly becuse I am "better at the game" witch i have a hard time interpreting. So other than rubbing my opponent of their Sv rolls, forcing morale fail and mortal wound bomb them what can I do to eliminate my opponents units reliably when probability keeps failing you? Are there better ways to mitigate probability?


Dude, wat?

"Backfield? I have no backfield." 
   
Made in se
Been Around the Block





 astro_nomicon wrote:
Gaen wrote:
Probability or statistics when you build your lists? Never been one to have much success going by probability and what I could potentially achieve as others often suggests to me. Statistically I get all the 1s when rolling to hit and my opponent get all the 6s when rolling to save, add all our rolls together and math for probability checks out on the whole. Solution, higher AP weapon to not allow the opponent to roll for save when I do get hits through. Works decently untill i meet an opponent with lots of inv and other counters. Trying to counter it wiht lots of mortal wounds but so fore its been unreliable for the most part. Thinking of trying out some Ld debuffs aswell to get more out of my successes. Rerolls are not a safe bet as i have had sets of 1s rerolled with a new set of dice shaking it up properly and ending upp more 1s. Though I have been told that I am rather obnoxious to play against no matter what list or army I use. Presumebly becuse I am "better at the game" witch i have a hard time interpreting. So other than rubbing my opponent of their Sv rolls, forcing morale fail and mortal wound bomb them what can I do to eliminate my opponents units reliably when probability keeps failing you? Are there better ways to mitigate probability?


Dude, wat?
Tldr, I suck att dice rolls what are good ways to be less reliable on the rng.

Deathwatch +3000p
Farsight +2000p
Kraken +2000p
Nephrekh +1000 
   
Made in nl
Dakka Veteran






To be honest, it seems very unlikely to me that some people are better or worse at dice than others. Of course, it's quite logical to see oneself as being less 'lucky' than others, almost everyone believes that. I've met dozens of people claiming to always have bad rolls and no one who claims to just have good rolls all the time. It's just that when things go wrong, one tends to notice the effect of the dice while when things go right or better than expected, that aspect tends to be ignored (except by the opponent).

And your comments make me wonder, do you mention your dice rolling during games as well?

   
Made in gb
Decrepit Dakkanaut




UK

Your post seems a bit all over the place.

Gaen wrote:
Probability or statistics when you build your lists? Never been one to have much success going by probability and what I could potentially achieve as others often suggests to me. Statistically I get all the 1s when rolling to hit and my opponent get all the 6s when rolling to save, add all our rolls together and math for probability checks out on the whole. Solution, higher AP weapon to not allow the opponent to roll for save when I do get hits through. Works decently untill i meet an opponent with lots of inv and other counters. Trying to counter it wiht lots of mortal wounds but so fore its been unreliable for the most part. Thinking of trying out some Ld debuffs aswell to get more out of my successes. Rerolls are not a safe bet as i have had sets of 1s rerolled with a new set of dice shaking it up properly and ending upp more 1s. Though I have been told that I am rather obnoxious to play against no matter what list or army I use. Presumebly becuse I am "better at the game" witch i have a hard time interpreting. So other than rubbing my opponent of their Sv rolls, forcing morale fail and mortal wound bomb them what can I do to eliminate my opponents units reliably when probability keeps failing you? Are there better ways to mitigate probability?


Probability is just probability. It doesn't care who you are. Now yes you can learn to trick-roll dice or use weighted dice to affect probability; but if you're normally rolling dice then chances are your rolls are landing within the realms of probabilities. However chances are that you don't remember the dice that roll average all the time. You remember the ones that roll worst or best and they stick in your mind far more so. That game where your anti-tank unit fails to kill a tank because of bad dice rolls - you remember it. However the other dozen times it does its average damage you tend to overlook or more easily forget because its not as big an event.
You can't mitigate probability without, basically, cheating. You can only work with it to maximise your predictions of what units might do whilst at the same time remaining flexible in your game thinking. Ergo don't rely on killing anything or saving any rolls; instead have a plan of attack, but be prepared at every turn to be reactive to a changing situation.

Gaen wrote:
Though I have been told that I am rather obnoxious to play against no matter what list or army I use. Presumebly becuse I am "better at the game" witch i have a hard time interpreting.


This is a bit worrying and confusing at the same time. You say that others consider you "obnoxious" to play against. You say that you presume its because you're better than them, however the rest of your post sounds like you're making losses because of bad dice rolls which would suggest you're losing more than your'e winning. So those two statements seem to be in conflict with each other. Furthermore winning lots doesn't make a person obnoxious to play against. At least by most definitions and use of the word that I'd associate it with "obnoxious" would be a term more reserved for someone who is socially bad to play against no matter if they win or lose. You might wish to talk with your opponents in a calm manner and ask them why they dislike playing against you. It most likely might have nothing to do with winning/losing/dice rolls and more to do with your behaviour and attitude whilst playing. Don't guess, ASK them. Otherwise its just a festering problem that will eventually end up with you not having people to play against, whilst being unable to correct the problem because you won't really know what the cause is.

A Blog in Miniature

3D Printing, hobbying and model fun! 
   
Made in us
Abel





Washington State

Probability = what might happen

Statistics = what actually happened

Kara Sloan shoots through Time and Design Space for a Negative Play Experience  
   
Made in de
Ladies Love the Vibro-Cannon Operator






Hamburg

Well, the more dice you are rolling, the better.
E.g., I prefer shuricannons over any other heavy weapon in the Eldar arsenal.
Combined with BT, I can reroll 1's and move and shoot without penalty (as the shuricannons are assault).
One-shot weapons usually suffer from the melta syndrome.
I stay away from them asap.

Former moderator 40kOnline

Lanchester's square law - please obey in list building!

Illumini: "And thank you for not finishing your post with a "" I'm sorry, but after 7200 's that has to be the most annoying sign-off ever."

Armies: Eldar, Necrons, Blood Angels, Grey Knights; World Eaters (30k); Bloodbound; Cryx, Circle, Cyriss 
   
Made in fr
Inquisitorial Keeper of the Xenobanks





France

Buy new dice ?

This message was edited 1 time. Last update was at 2020/04/03 10:58:09


   
Made in se
Been Around the Block





Trust me I have counted and that was before it became a runing joke in the group the certain people always rolled 1s to hit and other 6s on saves.

Also tried more dice only difference was how much more my opponent laugh when it became more apparent how often I rolled 1s.

Deathwatch +3000p
Farsight +2000p
Kraken +2000p
Nephrekh +1000 
   
Made in de
Ladies Love the Vibro-Cannon Operator






Hamburg

Gaen wrote:
Trust me I have counted and that was before it became a runing joke in the group the certain people always rolled 1s to hit and other 6s on saves.

Also tried more dice only difference was how much more my opponent laugh when it became more apparent how often I rolled 1s.

Well, in a whole game of 5 to 7 rounds, the law of large numbers should average out luck and bad luck when rolling dice.
Once I played a tourney with my Necrons in 6th edition and I failed almost every saving throw with my Wraiths in round 1.
Then I came back in round 2 and tabled his CSM army in round 5.

Former moderator 40kOnline

Lanchester's square law - please obey in list building!

Illumini: "And thank you for not finishing your post with a "" I'm sorry, but after 7200 's that has to be the most annoying sign-off ever."

Armies: Eldar, Necrons, Blood Angels, Grey Knights; World Eaters (30k); Bloodbound; Cryx, Circle, Cyriss 
   
Made in us
Longtime Dakkanaut




The best way to mitigate probability is usually stacking the odds in your favour - which is largely, but not entirely, the basis of list building. Its then playing so you throw mortal wounds into models with invuls, high AP attacks into things with good armour saves and so on. Avoid relying on low probability things.

But yeah, short of rigged dice or learning to roll them, luck is a feature of the game and there isn't much you can do beyond this. Sometimes you will be unlucky.

Its commonly said - but Bloodbowl was good for learning something about probability. The odds of rolling a double skull on a block seem low at 1/36 - but if you are a bashy team doing say 6-8 blocks a turn (okay your opponent probably shouldn't let you, but...) you'd expect it to happen once or twice a half. Then the dreaded double skulls into double skulls - 1/1296. Surely that's never going to happen? Well.. again, over the course of a few games, as you perform hundreds of blocks, the maths starts to become that yes, you'd expect to see it happen - and so when it does, don't be horrified. (Similar in a way for dodgy teams, although the numbers are a bit different.)
   
Made in jp
Unshakeable Grey Knight Land Raider Pilot




Is this meant to be a joke thread or something?
   
Made in us
Longtime Dakkanaut





Ellicott City, MD

Selective memory. Confirmation bias. Assuming data is the plural of anecdote. Please don’t go to Las Vegas. They make a very good living off of people who post threads like this...

Valete,

JohnS

Valete,

JohnS

"You don't believe data - you test data. If I could put my finger on the moment we genuinely <expletive deleted> ourselves, it was the moment we decided that data was something you could use words like believe or disbelieve around"

-Jamie Sanderson 
   
Made in us
Wicked Canoptek Wraith



United States

Buy some casino dice. They're more uniform and consistent in their rolls. Dice with rounded edges yield different results statistically versus dice with flat/squared edges.
   
Made in us
Pious Palatine




punisher357 wrote:
Buy some casino dice. They're more uniform and consistent in their rolls. Dice with rounded edges yield different results statistically versus dice with flat/squared edges.


Casino dice are terrible for stuff like this unless you're planning to whip them down a craps table though. In fact, the only reason I would bother with casino dice in 40k is because them being square makes them super easy to cheat with.


 
   
Made in us
The Marine Standing Behind Marneus Calgar





Upstate, New York

 Overread wrote:

Probability is just probability. It doesn't care who you are. Now yes you can learn to trick-roll dice or use weighted dice to affect probability; but if you're normally rolling dice then chances are your rolls are landing within the realms of probabilities. However chances are that you don't remember the dice that roll average all the time. You remember the ones that roll worst or best and they stick in your mind far more so. That game where your anti-tank unit fails to kill a tank because of bad dice rolls - you remember it. However the other dozen times it does its average damage you tend to overlook or more easily forget because its not as big an event.
You can't mitigate probability without, basically, cheating. You can only work with it to maximise your predictions of what units might do whilst at the same time remaining flexible in your game thinking. Ergo don't rely on killing anything or saving any rolls; instead have a plan of attack, but be prepared at every turn to be reactive to a changing situation.


I want to address the bolded bit. (fully agree with you on perception vs. probability)

There are ways in listbuilding to minimize chance. Take weapons that do a flat 3 damage over a d6. Avoid guns with random shots. Look for flat bonuses rather then variable. Ways to ignore dice rolls/fix results.

Once you are flinging our fickle little 6-sided friends around, you are in Fate’s hands. So do what you can to avoid that.

   
Made in us
Consigned to the Grim Darkness





USA

Probability and statistics are the same thing. Don't conflate your outlier past results with future results.

The people in the past who convinced themselves to do unspeakable things were no less human than you or I. They made their decisions; the only thing that prevents history from repeating itself is making different ones.
-- Adam Serwer
My blog
 
   
Made in us
Fixture of Dakka






I had a game i quite 1/2 through my turn 2 b.c of bad dice rolls, but at the same time i have had insane dice rolls. Thats how statistics works with independent roles like dice.

20 roles of the dice isn't 1 stat with 20 results, its 20 single pieces of data, in theory you can role 20 1's in a row b.c each role is independent of the other. Yes when looking at the over all statistics thats out side of the normal, but its still able to do that.

YJust like when the same person wins the lottery more than once.

   
Made in us
Wicked Canoptek Wraith



United States

ERJAK wrote:
punisher357 wrote:
Buy some casino dice. They're more uniform and consistent in their rolls. Dice with rounded edges yield different results statistically versus dice with flat/squared edges.


Casino dice are terrible for stuff like this unless you're planning to whip them down a craps table though. In fact, the only reason I would bother with casino dice in 40k is because them being square makes them super easy to cheat with.


So Casinos use dice that are easy for the person rolling to cheat with? Riiiiiggghhht. Square, sharp edged dice are easier to cheat with? Sounds baseless to me.

It's fine if you don't like them, but they do roll more consistently. There are plenty of articles and studies you can read about online. I'm not going to bother listing them.
   
Made in es
Guard Heavy Weapon Crewman




Gaen wrote:
Though I have been told that I am rather obnoxious to play against no matter what list or army I use. Presumebly becuse I am "better at the game" witch i have a hard time interpreting.

So you are either trolling or that guy being self righteous right now, the only thing i can answer to this post “git gud” XD.


Automatically Appended Next Post:
punisher357 wrote:
ERJAK wrote:
punisher357 wrote:
Buy some casino dice. They're more uniform and consistent in their rolls. Dice with rounded edges yield different results statistically versus dice with flat/squared edges.


Casino dice are terrible for stuff like this unless you're planning to whip them down a craps table though. In fact, the only reason I would bother with casino dice in 40k is because them being square makes them super easy to cheat with.


So Casinos use dice that are easy for the person rolling to cheat with? Riiiiiggghhht. Square, sharp edged dice are easier to cheat with? Sounds baseless to me.

It's fine if you don't like them, but they do roll more consistently. There are plenty of articles and studies you can read about online. I'm not going to bother listing them.

Casinos also require you to throw in a specific manner for a reason.

This message was edited 2 times. Last update was at 2020/04/30 16:52:48


 
   
Made in gb
Longtime Dakkanaut




punisher357 wrote:
ERJAK wrote:
punisher357 wrote:
Buy some casino dice. They're more uniform and consistent in their rolls. Dice with rounded edges yield different results statistically versus dice with flat/squared edges.


Casino dice are terrible for stuff like this unless you're planning to whip them down a craps table though. In fact, the only reason I would bother with casino dice in 40k is because them being square makes them super easy to cheat with.


So Casinos use dice that are easy for the person rolling to cheat with? Riiiiiggghhht. Square, sharp edged dice are easier to cheat with? Sounds baseless to me.

It's fine if you don't like them, but they do roll more consistently. There are plenty of articles and studies you can read about online. I'm not going to bother listing them.


Casinos require the dice to be thrown and bounce off the wall at the end of a craps table, which is what imparts true randomness. Casino dice without that caveat are actually among the easiest to manipulate exactly because they are so precisely balanced. Also, casino dice suck for 40k because they're far too big to be practical.

As far as the OP goes, this is a joke, right? If it isn't then no, you're not any more or less lucky than your opponents unless your dice are actually loaded or weighted towards certain results, either intentionally or not. 40k list building is essentially about mitigating luck as much as possible, which is why so many successful armies contain loads of re-rolls and bonuses to hit.
   
Made in de
Toothy 3rd Gen True Hybrid





Someone invented a perfect set of houserules to avoid all matters of luck.
You just need to follow a very specific setup for your armies and a specially prepared gaming table.

Hmm, maybe I can find a link to that homebrew...

Ahh, here it is https://www.officialgamerules.org/chess

This message was edited 1 time. Last update was at 2020/04/30 18:44:59


   
Made in us
Heroic Senior Officer





Western Kentucky

My only real input to this is plan for the worst, hope for the best (or in this case, average) main thing I do is round "down" (aka bad for me) as a rule anytime I get a decimal for the average. Yes, the average on a d6 is 3.5 for example, but you can't roll a 3.5. So anytime I do math involving a d6 when I'm considering running something, I go ahead and round down to 3. However, if I'm rolling 2d6, then it makes sense to take the average of 7, because odds are one would be 3 and one would be 4, but I often find myself still banking on 6. Conversely, if it's the opponent, I round in their favor. So for example, if my Russ is being shot at by a lascannon, I expect it to do 4 damage, even though average is 3.5. This is important because this means I expect a Russ to die to 3 lascannon wounds getting through, not 4, which is what the average would tell you is the case. This means that if the opponent has a little bit of a hot streak on their dice, I'm already prepared for that eventuality as best I can and have already considered what to do now.

This has nothing to do with statistics or math and all down to common sense and being prepared. Averages are good for getting ballpark estimates and roughly determining what a unit will do in a given situation, but one should never count on averages as a set in stone rule. The odds of every roll in a game being the average you plan on are possibly just as low as rolling nothing but 1's. You should have backup plans built in for units rolling poorly or an opponent having a hotstreak. The dice gods are fickle gods after all. This is the kind of thing where you go from mathhammer and theory hammer to actual tactics and practice. Granted the kind of work I do we commonly work with unknown factors and gut decisions based on preparing for the worst, so that mindset has crept into 40k, but that's not to say it doesn't have value here. Statistics and averages are incredibly useful and important tools, but unless you're prenerf iron hands or something, most games will still be decided through a combination of strategy, tactics, and pre-planning (mathhammer and tactics), not just solely averages through math hammer.

Simply put, planning for the worst and hoping for the best means that worst case scenario, you're prepared, and anything in your favor is just gravy and a happy accident. At that point there are way less "surprises" and it helps keep you from being tilted as much after an opponent has a hot streak on their dice if you planned on them being a bit lucky and you being a bit unlucky to begin with. Hopefully that makes sense. I'm not saying that averages aren't incredibly important or luck isn't a thing, just that I plan on being a bit unlucky and that's helped a lot.

'I've played Guard for years, and the best piece of advice is to always utilize the Guard's best special rule: "we roll more dice than you" ' - stormleader

"Sector Imperialis: 25mm and 40mm Round Bases (40+20) 26€ (Including 32 skulls for basing) " GW design philosophy in a nutshell  
   
Made in us
Longtime Dakkanaut





Ellicott City, MD

 AuntHerbert wrote:
Someone invented a perfect set of houserules to avoid all matters of luck.
You just need to follow a very specific setup for your armies and a specially prepared gaming table.

Hmm, maybe I can find a link to that homebrew...

Ahh, here it is https://www.officialgamerules.org/chess


Bah... White has a significant first turn advantage in Chess.. It’s anything but free of luck. Picking the lett hand or right to determine who is White and who is Black is hardly removing luck!



Valete,

JohnS

Valete,

JohnS

"You don't believe data - you test data. If I could put my finger on the moment we genuinely <expletive deleted> ourselves, it was the moment we decided that data was something you could use words like believe or disbelieve around"

-Jamie Sanderson 
   
Made in us
Locked in the Tower of Amareo




The false equivalence of chess is always baffling.
   
Made in us
Consigned to the Grim Darkness





USA

Also, chess sucks and lacks creativity.

The people in the past who convinced themselves to do unspeakable things were no less human than you or I. They made their decisions; the only thing that prevents history from repeating itself is making different ones.
-- Adam Serwer
My blog
 
   
Made in us
Daemonic Dreadnought





Eye of Terror

 cygnnus wrote:
Bah... White has a significant first turn advantage in Chess.. It’s anything but free of luck. Picking the lett hand or right to determine who is White and who is Black is hardly removing luck!



Valete,

JohnS

Former competitive chess player here. First turn advantage in chess is a myth, should not be taken for granted.

The evidence for white supremacy typically falls into 3 categories:

- Personal experience of Chess masters
- Aggregate games from various forums
- Computer simulations

While each category will demonstrate a bias towards white, each category has flaws. Books by chess masters are anecdotal and often controversial, aggregate game studies almost always fail to account for differences in relative skill level and assume the "better" player randomly selects between white and black, and computer simulations fail to account for outcome bias in playstyle (i.e. computers will play for a draw in situation where a human would play for a win.)

One could argue it's probable that white will win, based on historical outcomes, opinions, and computer performance.

One could also argue it's more likely black will win based on better statistical ranking of the black player, aversion to seeking a draw, and limitations to technology (i.e. computers still haven't solved chess.)

If you look at chess as a game of moving pieces on a board, sure, white has an advantage because it goes first. If you look at chess as a game where two people compete to achieve a certain outcome, and preparation / motivations / independence from previous trials matter, white has no advantage. If you don't believe the later, look at Teimour Radjabov.

   
Made in de
Toothy 3rd Gen True Hybrid





Martel732 wrote:
The false equivalence of chess is always baffling.


The probability, that someone who drops the term "false equivalence" doesn't even understand what it means, is even more breathtaking.

   
 
Forum Index » 40K General Discussion
Go to: