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.

Hey folks!

Long time absence from 40k, freshly returning! As you read this, I'd like you to consider two questions.


-What heuristic weighting scores do you use in the framework of the equation below?
-What if any variables do you believe are missing from this equation?

I posit that the competitive capabilities of a player at a table can be represented by an equation.

Let's say that this equation is X = A + B + C, where heuristic coefficients totaling (1) weight each variable, and the variables are defined thus:

X is the competitiveness of a player's army in a tournament
A is the player's tactical prowess
B is the strength of a codex in a vacuum
C is the die results on the table.

My heuristic coefficients - based on subjective experiences like any heuristic weighting mechanism - probably look like this:
X = 0.75A + .15B + .1C.

What I feel like I'm reading now - and long ago seems to look like one of these two:
X = .25A + .5B + .25C
OR
X = .2A + .6B + .2C


I could be wrong; but every informed decision requires a body of empirical evidence to support it, so I am gathering information from you! Since this is a heuristic equation, there is no right or wrong. Your weighting of coefficients is equally valid to mine.

I don't know what my weighting coefficients ACTUALLY are for this equation, except insofar that I believe A > B+C by such a large margin that B and C are negligible to the overall valuation of the equation. I am assuming (and welcoming challenges to my assumption) that since most people cannot control for A, they overvalue B and C. This is a heuristic equation, you can't win or lose. There is no right or wrong, only hearing other interpretations of subjective math.

With that said, I'm very interested to hear your math on competitive weighting.

Depends on the situation. Talking about all players? Almost 100% B, a good tournament list is going to destroy a newbie's first battleforce army with even moderate player skill. But if you limit it to the top 25% at major tournaments, where it's assumed that everyone has at least a decent list, then A and C become a lot more important.

I don't know if I have a direct answer, but I have a handful of thoughts on the matter:

- Ultimately almost all outcomes are decided by die rolls, so as any game progresses beyond the start, die rolls become increasingly relevant until you hit the die roll that determines the winner of the game.

This isn't necessarily the last roll of the game, but that roll that determines whether or not one person is effectively removed from having any possibility of winning still. It could be the last armor save roll on a pivotal character, a charge roll, or that classic "I need to run 5 inches onto the objective otherwise I lose the game" scenario. You can argue that those things can be mitigated by tactical acumen, and to a degree they can, but at the end of the day even the most well planned out play fails if you roll nothing but ones.

- There should either be another factor including a given opponents list/codex comp, or that should be explicitly taken into account in B.

My argument for this is that, in almost any codex, you can build a list that will hard counter another list, but will be almost utterly ineffective against it's hard counter. Plasma spam is great against a hypothetical elite army like GK (sidestepping that most things are), but it's not going to do nearly as well as massed small arms will against a guard horde. There's an undeniable meta element present in list building.

- Variable A for any given player may be different when using a given codex than it would be with a different one.

I'm a pretty good guard player and not half bad with SM. I CAN play Nids, but I am not nearly so good there. I'm willing to bet (especially in these brave days of imperial soup!) that many people have multiple armies and inconsistent skill levels between them all.

I agree with Peregrine on A and C being significantly more important in places where B is going to be consistently high. I would go one step further and say that as A becomes consistently high, the thing that matters most at all is the die rolls.

I'd guess casual is probably about the .25A + .5B + .25C estimate.

I would say that a highly competitive environment (LVO or Adepticon level) is probably about .3A + .1B + .6C.

 daedalus wrote:
- Ultimately almost all outcomes are decided by die rolls, so as any game progresses beyond the start, die rolls become increasingly relevant until you hit the die roll that determines the winner of the game.

This isn't necessarily the last roll of the game, but that roll that determines whether or not one person is effectively removed from having any possibility of winning still. It could be the last armor save roll on a pivotal character, a charge roll, or that classic "I need to run 5 inches onto the objective otherwise I lose the game" scenario. You can argue that those things can be mitigated by tactical acumen, and to a degree they can, but at the end of the day even the most well planned out play fails if you roll nothing but ones.

Disagree on two points here:

1) Dice are IMO much less of a factor than in other games because of how many you roll. Over the length of a full game you're going to get close to that bell curve distribution for both players, and the game is designed around rolling buckets of dice with each die having a low individual probability of success. It's not like, say, X-Wing where you only have 3-4 units on each side and a single roll of 3 dice can decide if you win or lose.

2) You're focusing too much on single results because they happened to be at the very end and ignoring the context. For example, if the last failed armor save on a character effectively ends the game then you're probably in such a poor position that you were going to lose anyway, possibly on the very next die roll when you failed that one instead. 40k is such a big game that the contribution of any individual event is small, and it's rare that a single roll of the dice is deciding everything. It's just tempting to focus on the single event because it's an explanation that feels good ("I lost because I couldn't roll a 5+) instead of the more difficult question of how you got yourself into a position where your only hope of winning came down to a single roll of the dice.

This is an interesting question I think that values are something like this in a serious tournament:

X = (0.5)A + (0.05)B + (0.1)C + (0.1)D + (0.25)E

A is important, and I'm assuming by "tactical prowess" you are excluding list construction.
B is very low since in a tournament there will be few people playing weak codexes or playing weak units from codexes.
C is low due to the amount of dice rolled and the ability of a smart player to plan around dice variance.

D and E are for other. Your equation does not include die/random results off of the table, like which players are matched up in the tournament rounds which I put into D. It also does not include list construction and/or army composition, which I put into E.

I interpret "the competitiveness of a player's army in a tournament " as the chance of the player + army placing highly in the tournament. If a player+army is matched against a skilled player with a good army early in the tournament, then the chance of that player+army placing highly will drop. Additionally, it is hard to make a truly "Take All Comers" list, so being matched against an army that your army is weak against lowers your chances of placing highly as well.

P.S.: I have a feeling what I have for B is not what you meant by "the strength of a codex in a vacuum" since that does not imply anything about the codex(es) the player is actually using. For example, if there was one god codex that everyone played at a competitive level then B would be 0 since there are no differences at a competitive level between the codexes being played. If you meant something more like "the player's army composition", then add B and E together.

Think B might need some further pondering, as on paper the poster boy Marines seem good, decent armour, nice range of guns and toys, their shortcomings only come to light when its clear other armys can negate a lot of their advantages via weight of dice, mortal wounds and even bigger guns

It doesn't even have to stay in the format I posited in the OP. A sports bookie lays odds on a game based on a statistical evaluation of...THINGS. Those variables can be weighted to fit a predictive model of accuracy. I'm trying to figure out how to express this problem in math.

If we were to reformulate this question into something like, "If you were a 40k bookie working out the odds in a fight, how would you create the odds?" And can the community agree on how the odds should be formulated? If there can be a common, mathematical understanding of the heuristics and the model, people can weight variables based on their own beliefs - which are relevant insofar that there is ample data available to create a suitable means of testing the accuracy of it.

If Player 1 plays (Army A) and Player 2 players (Army B), there must be a realistic means of expressing the likely statistical result of that game based on some variables known in advance.

 Peregrine wrote:


1) Dice are IMO much less of a factor than in other games because of how many you roll. Over the length of a full game you're going to get close to that bell curve distribution for both players, and the game is designed around rolling buckets of dice with each die having a low individual probability of success. It's not like, say, X-Wing where you only have 3-4 units on each side and a single roll of 3 dice can decide if you win or lose.

Depends on the army and point values I guess. Guard will smooth out much more than an elite army. I played 1000 points (Adepticon TT rules) of DA Primaris yesterday and probably threw less than 300 dice the whole game. Meanwhile, I've had 1000 points of guard that could probably throw that many dice in a round without breaking a sweat.

2) You're focusing too much on single results because they happened to be at the very end and ignoring the context. For example, if the last failed armor save on a character effectively ends the game then you're probably in such a poor position that you were going to lose anyway, possibly on the very next die roll when you failed that one instead. 40k is such a big game that the contribution of any individual event is small, and it's rare that a single roll of the dice is deciding everything. It's just tempting to focus on the single event because it's an explanation that feels good ("I lost because I couldn't roll a 5+) instead of the more difficult question of how you got yourself into a position where your only hope of winning came down to a single roll of the dice.

For the ones where you sit there and think "okay, this is it"? Yeah, I could agree with that usually. If I get to that point, it probably means I've done something wrong. I kind of consider the game a waste if I don't come away finding at least one thing I could have done better somewhere.

Having that been said, it's a game that has a linear progression of events with a diminishing economy of possibilities. Sooner or later there will be an event that decides the game, even if it's non-obvious which one it is. I guess it won't always be a SINGLE die roll, but it will almost always be something decided by dice.

 Dashofpepper wrote:
It doesn't even have to stay in the format I posited in the OP. A sports bookie lays odds on a game based on a statistical evaluation of...THINGS. Those variables can be weighted to fit a predictive model of accuracy. I'm trying to figure out how to express this problem in math.

If we were to reformulate this question into something like, "If you were a 40k bookie working out the odds in a fight, how would you create the odds?" And can the community agree on how the odds should be formulated? If there can be a common, mathematical understanding of the heuristics and the model, people can weight variables based on their own beliefs - which are relevant insofar that there is ample data available to create a suitable means of testing the accuracy of it.

If Player 1 plays (Army A) and Player 2 players (Army B), there must be a realistic means of expressing the likely statistical result of that game based on some variables known in advance.

there are so many other variables to take into account that it would become such a monster formulae that would get so unreliable as to be worthless.

As a thought exercise, I think we all wonder at times if it's the faction carrying the winner to victory or his skill with that faction and thus if we could replicate it by taking the same list.

but in reality you'd need to include variables like

time since last rules update
Terrain rules / table layout
who went first
deployment zone types
stratagems and spent CP
and once models hit the table there's a million billion different variations of where stuff can move to that getting the same players to play the same game 'move for move' would potentially yield different results.

Right down to did one guy go for a nervous poop before the match or not.

The best you could get is rough rankings - as posters above have said -

good list + good player
bad list + good player / good list + new player
bad list + bad player

and all you need to look at here is pick a faction that was popular from recent tournament X - i.e. 143 players took IK/BA/IG and look at the rankings ... and there you have it

no need for verbose formulae

Reanimation_Protocol wrote:


there are so many other variables to take into account that it would become such a monster formulae that would get so unreliable as to be worthless.

As a thought exercise, I think we all wonder at times if it's the faction carrying the winner to victory or his skill with that faction and thus if we could replicate it by taking the same list.

but in reality you'd need to include variables like

time since last rules update
Terrain rules / table layout
who went first
deployment zone types
stratagems and spent CP
and once models hit the table there's a million billion different variations of where stuff can move to that getting the same players to play the same game 'move for move' would potentially yield different results.

Right down to did one guy go for a nervous poop before the match or not.

The best you could get is rough rankings - as posters above have said -

good list + good player
bad list + good player / good list + new player
bad list + bad player

and all you need to look at here is pick a faction that was popular from recent tournament X - i.e. 143 players took IK/BA/IG and look at the rankings ... and there you have it

no need for verbose formulae

I don't agree with the concept that problem statements can be so complex that they cannot be expressed.

Reanimation_Protocol wrote:

and all you need to look at here is pick a faction that was popular from recent tournament X - i.e. 143 players took IK/BA/IG and look at the rankings ... and there you have it

I don't agree with that either. That presents armies in a vacuum, detached from the OTHER things that influence rankings. Consider this:

-In your example, 143 players took IK/BA/IG
-Let's say that the top 10 armies were all Imperial Knights.
-People therefore make the conclusion that Imperial Knights are the strongest army.


To make an extreme example, there are some universally accepted "good players" right? Reece Robbins, Nick Nanaveti, Hulksmash, Paul Murphy .... I've been gone a long time, I don't know who is good or not. Let's say it's those guys and they all play ....sisters of battle. I would expect the example to be modified thusly:

-143 players took IK/BA/IG. 4 Players brought SoB.
-The top 10 armies were 6 Imperial Knight armies and 4 Sisters of Battle armies.
-People therefore make the conclusion that Imperial Knights and Sisters of Battle are the strongest armies.

It is not easy to quantify heuristic metrics, but if such an exercise were to be completed and universally accepted, people would have a relative starting point to discuss the merits, strengths, and weaknesses of a given codex in a way that controls for OTHER variables. If Player 1 plays (Army A) and Player 2 players (Army B), there must be a realistic means of expressing the likely result of that game based on some variables known in advance, where such variables and their weighting mechanism can be commonly agreed upon.

An equation is simple, and not ideal - what I'm trying to visualize here is modeling the expected performance of controlled variables such that people could be modeled to take an identical list through a series of identical tasks with identical die results, then capture data. That lets you control for A. Etc, etc, creating data points to hone in an algorithm or modeling platform that would then be able to accurately forecast what will happen in a game before hand.

Every sport does it; statistical analysis to come up with trends. Those in turn create predictive models or something that bookies can use to create odds for people to bet against. Those odds are based on....I have no idea what they are based on. I guess I'm basically asking "If you were a bookie, what is your math and the variables for predicting the likely outcome of a game?" I took a stab at a simple formula to do it; I'd like to know how YOU do it.

The answer can't be, "It's too complex to calculate."

Reanimation_Protocol wrote:

 Dashofpepper wrote:
It doesn't even have to stay in the format I posited in the OP. A sports bookie lays odds on a game based on a statistical evaluation of...THINGS. Those variables can be weighted to fit a predictive model of accuracy. I'm trying to figure out how to express this problem in math.

If we were to reformulate this question into something like, "If you were a 40k bookie working out the odds in a fight, how would you create the odds?" And can the community agree on how the odds should be formulated? If there can be a common, mathematical understanding of the heuristics and the model, people can weight variables based on their own beliefs - which are relevant insofar that there is ample data available to create a suitable means of testing the accuracy of it.

If Player 1 plays (Army A) and Player 2 players (Army B), there must be a realistic means of expressing the likely statistical result of that game based on some variables known in advance.

there are so many other variables to take into account that it would become such a monster formulae that would get so unreliable as to be worthless.

As a thought exercise, I think we all wonder at times if it's the faction carrying the winner to victory or his skill with that faction and thus if we could replicate it by taking the same list.

but in reality you'd need to include variables like

time since last rules update
Terrain rules / table layout
who went first
deployment zone types
stratagems and spent CP
and once models hit the table there's a million billion different variations of where stuff can move to that getting the same players to play the same game 'move for move' would potentially yield different results.

Right down to did one guy go for a nervous poop before the match or not.

The best you could get is rough rankings - as posters above have said -

good list + good player
bad list + good player / good list + new player
bad list + bad player

and all you need to look at here is pick a faction that was popular from recent tournament X - i.e. 143 players took IK/BA/IG and look at the rankings ... and there you have it

no need for verbose formulae

A lot of those factors come down to the tactical prowess of the player, though, which is A in the OP's formula. Some of them elements of the equation would likely be interdependent, I think. For example, good players often build lists and play in such a way to mitigate against the randomness of the dice, so it could be argued that the better the player, the less weighting should be applied to C in the original equation. Some factions are very good at getting rerolls too, so the relevance of C could depend very much on B. Having said that, I think the final part in the post quoted above is probably as good as any more complicated formula.

Slipspace wrote:


A lot of those factors come down to the tactical prowess of the player, though, which is A in the OP's formula. Some of them elements of the equation would likely be interdependent, I think. For example, good players often build lists and play in such a way to mitigate against the randomness of the dice, so it could be argued that the better the player, the less weighting should be applied to C in the original equation. Some factions are very good at getting rerolls too, so the relevance of C could depend very much on B. Having said that, I think the final part in the post quoted above is probably as good as any more complicated formula.

I've seen a lot of discussions over the years where people make the argument that 40k is a luck-based game determined by dice. I think that part of player skill is to control for the randomness of dice by mitigating the important of dice; movement, positioning, cover, proximity, target confusion, or even using dice quantity instead of quality as a means of trying to control probability expectations.

So for me, if "die results" are expressed as variable "C" I weight C in that overall equation as ...maybe .1C. Other people think dice are much more important and weight it as .5C.

I'm not trying to make everyone agree on how to weight those variables with coefficients, I'm just trying to accurately capture what the universally understood variables ARE to start with.

 Peregrine wrote:

1) Dice are IMO much less of a factor than in other games because of how many you roll. Over the length of a full game you're going to get close to that bell curve distribution for both players, and the game is designed around rolling buckets of dice with each die having a low individual probability of success. It's not like, say, X-Wing where you only have 3-4 units on each side and a single roll of 3 dice can decide if you win or lose.

This depends on how true the dice are. It's not hard to see that most GW dice aren't even close to uniformly shaped/weighted. Heck, look at those squig dice, those irregular faces are NOT going to roll true!

This may not be too complex to calculate, but given the number of variables, I'm concerned that we're dealing with too small of a sample size for anything other than first order approximations.

 Peregrine wrote:

1) Dice are IMO much less of a factor than in other games because of how many you roll. Over the length of a full game you're going to get close to that bell curve distribution for both players, and the game is designed around rolling buckets of dice with each die having a low individual probability of success. It's not like, say, X-Wing where you only have 3-4 units on each side and a single roll of 3 dice can decide if you win or lose.

This is probably the most often perpetutated myth on dakka and shows total lack of understanding how probabilities and statistics interact with eachother in games...

Large number of dice equalizes single event outcomes - like guard blob shooting, but it doesn't do anything for chains of causality connected events that constitute a game of 40K. In other words - non-average rolls of large variability events weight a lot more in early turns and on actions taken early (they cut down game history branches that are closer to root), even if common perception of "deciding roll" is that it happens late in game. Even if you consider things like single round of shooting of 5 quads of 5 tacticals, the outcome of such round cannot be reasonably approximated by averages, because single events that roll small number of dice have much higher variance of results than a single large squad of 25 tacticals shooting together, and you base your decisions about what 2nd, 3rd, 4th and 5th squad of marines does on the outcomes of previous events. This is really probability and statistics 101 and the reason why low discrepancy weapons/units are better competetive choices than theoretically powerfull units with huge swings in the damage output. You can however judge the overall performance of a given codex unit via "theoryhammer", because that is reasoning that is causal independet and will even itself over a course of many whole games.

And what Daedalus wrote is true - in every game of 40K ever played there is a point at which the winner is already decided, even if players are not aware of this fact yet, because every game, even luck-based, can be analysed by decision tree, and within each tree there is a point, in which remaining decision branches are too short to change the outcome. his may be the "who goes first" roll in case of badly biased matchup or it can be a single close roll in the last turn of the last round. The overall good casual game design targets deciding rolls to be as close to the end of the game as possible, but in sandbox games (and 40K is a sandbox game) it is literally impossible to achieve this, so in result most 40K games are decided rather earlier than later into a match (and no, X-Wing considered as a whole system, including all unusable cards and ships is no better than 40K in that regard, the main reason why it is perceived as ballanced better is that x-wing is much closer to CCG in terms of emotional/time involvement into collecting/changing lists, so it is far easier for players to chase the meta, so in turn it is much less of a sandbox game). This is also why many games introduce tiered win classes, because even after this deciding point in a game, there are still decisions remaning that can alter the exact discrepancy of players scores even if they cannot change the winner.


Automatically Appended Next Post:
And a little thought experiment to chew on - the next time you play, record the total number of dice rolled throughout the game (or you can even record actual results), and the next time you play, allow both players to utilise a pool of dice results - either experimentally collected or prepared based on simple division of entire pool into six categories of results. Such pool will, by definition, follow average distribution exactly. The resulting game of 40K will be an utter abomination and won't resemble typical match in the slightest.

 daedalus wrote:
Depends on the army and point values I guess. Guard will smooth out much more than an elite army. I played 1000 points (Adepticon TT rules) of DA Primaris yesterday and probably threw less than 300 dice the whole game. Meanwhile, I've had 1000 points of guard that could probably throw that many dice in a round without breaking a sweat.

That's just proving my point. You played a small game (half the typical point level) with an elite army and you still rolled 300 dice. A game of X-Wing is typically decided by somewhere in the range of 10-20 dice.

Having that been said, it's a game that has a linear progression of events with a diminishing economy of possibilities. Sooner or later there will be an event that decides the game, even if it's non-obvious which one it is. I guess it won't always be a SINGLE die roll, but it will almost always be something decided by dice.

Again, you're focusing on a single event as being THE event and ignoring all of the other events that would have happened if it had gone differently. Let's imagine an extreme example: a game with a 2000 point optimized tournament list against an IG patrol detachment consisting of a single company commander and a single infantry squad with no upgrades at all. There will be a single event that decides the game (probably in the first shooting phase), but even if that particular event hadn't been the deciding one the player with a real list has such an overwhelming advantage that even if that specific event hadn't decided the game it would just have been decided by a different event a few moments later. The same kind of thing usually happens in normal games, the outcome is the sum of many smaller events that each contribute to giving or taking advantage from one player or the other but there's rarely a single one that decisively swings it.

A better way to look at it is to think about win probability graphs in sports. At every moment of the game you plot the expected win probability (based on score, position on the field, etc) for each team. You'll see that it changes over the course of the game, but no single event permanently sets it to a win for either team. You're looking at a trend line, not a discrete point, and even "going to win" is determined by asking when a team's win probability permanently crossed a certain arbitrary line where a comeback by the loser was low enough in probability rather than a definite yes/no answer.

nou wrote:

 Peregrine wrote:

1) Dice are IMO much less of a factor than in other games because of how many you roll. Over the length of a full game you're going to get close to that bell curve distribution for both players, and the game is designed around rolling buckets of dice with each die having a low individual probability of success. It's not like, say, X-Wing where you only have 3-4 units on each side and a single roll of 3 dice can decide if you win or lose.

This is probably the most often perpetutated myth on dakka and shows total lack of understanding how probabilities and statistics interact with eachother in games...

Large number of dice equalizes single event outcomes - like guard blob shooting, but it doesn't do anything for chains of causality connected events that constitute a game of 40K. In other words - non-average rolls of large variability events weight a lot more in early turns and on actions taken early (they cut down game history branches that are closer to root), even if common perception of "deciding roll" is that it happens late in game. Even if you consider things like single round of shooting of 5 quads of 5 tacticals, the outcome of such round cannot be reasonably approximated by averages, because single events that roll small number of dice have much higher variance of results than a single large squad of 25 tacticals shooting together, and you base your decisions about what 2nd, 3rd, 4th and 5th squad of marines does on the outcomes of previous events. This is really probability and statistics 101 and the reason why low discrepancy weapons/units are better competetive choices than theoretically powerfull units with huge swings in the damage output. You can however judge the overall performance of a given codex unit via "theoryhammer", because that is reasoning that is causal independet and will even itself over a course of many whole games.

And what Daedalus wrote is true - in every game of 40K ever played there is a point at which the winner is already decided, even if players are not aware of this fact yet, because every game, even luck-based, can be analysed by decision tree, and within each tree there is a point, in which remaining decision branches are too short to change the outcome. his may be the "who goes first" roll in case of badly biased matchup or it can be a single close roll in the last turn of the last round. The overall good casual game design targets deciding rolls to be as close to the end of the game as possible, but in sandbox games (and 40K is a sandbox game) it is literally impossible to achieve this, so in result most 40K games are decided rather earlier than later into a match (and no, X-Wing considered as a whole system, including all unusable cards and ships is no better than 40K in that regard, the main reason why it is perceived as ballanced better is that x-wing is much closer to CCG in terms of emotional/time involvement into collecting/changing lists, so it is far easier for players to chase the meta, so in turn it is much less of a sandbox game). This is also why many games introduce tiered win classes, because even after this deciding point in a game, there are still decisions remaning that can alter the exact discrepancy of players scores even if they cannot change the winner.


Automatically Appended Next Post:
And a little thought experiment to chew on - the next time you play, record the total number of dice rolled throughout the game (or you can even record actual results), and the next time you play, allow both players to utilise a pool of dice results - either experimentally collected or prepared based on simple division of entire pool into six categories of results. Such pool will, by definition, follow average distribution exactly. The resulting game of 40K will be an utter abomination and won't resemble typical match in the slightest.

You are correct - but we can't apply data points into a model to test accuracy without an algorithm to test, which in turn require variables, which brings us back to the root question of "How can we express the variables that make up the competitiveness of a given situation in a logical form?"

Take a stab at it!

 Dashofpepper wrote:
The answer can't be, "It's too complex to calculate."

Of course it can be. You personally might not like hearing the truth, but that doesn't make it any less true if the reality of the situation is in fact that it's too complex to calculate an expected winner between two reasonably comparable players/armies.

 Peregrine wrote:

 daedalus wrote:
Depends on the army and point values I guess. Guard will smooth out much more than an elite army. I played 1000 points (Adepticon TT rules) of DA Primaris yesterday and probably threw less than 300 dice the whole game. Meanwhile, I've had 1000 points of guard that could probably throw that many dice in a round without breaking a sweat.

That's just proving my point. You played a small game (half the typical point level) with an elite army and you still rolled 300 dice. A game of X-Wing is typically decided by somewhere in the range of 10-20 dice.

Having that been said, it's a game that has a linear progression of events with a diminishing economy of possibilities. Sooner or later there will be an event that decides the game, even if it's non-obvious which one it is. I guess it won't always be a SINGLE die roll, but it will almost always be something decided by dice.

Again, you're focusing on a single event as being THE event and ignoring all of the other events that would have happened if it had gone differently. Let's imagine an extreme example: a game with a 2000 point optimized tournament list against an IG patrol detachment consisting of a single company commander and a single infantry squad with no upgrades at all. There will be a single event that decides the game (probably in the first shooting phase), but even if that particular event hadn't been the deciding one the player with a real list has such an overwhelming advantage that even if that specific event hadn't decided the game it would just have been decided by a different event a few moments later. The same kind of thing usually happens in normal games, the outcome is the sum of many smaller events that each contribute to giving or taking advantage from one player or the other but there's rarely a single one that decisively swings it.

A better way to look at it is to think about win probability graphs in sports. At every moment of the game you plot the expected win probability (based on score, position on the field, etc) for each team. You'll see that it changes over the course of the game, but no single event permanently sets it to a win for either team. You're looking at a trend line, not a discrete point, and even "going to win" is determined by asking when a team's win probability permanently crossed a certain arbitrary line where a comeback by the loser was low enough in probability rather than a definite yes/no answer.

I would like to remind you here, that you are the one who constantly argues how for at least 3 editions, games of 40K are definitely decided at most at the "who goes first" roll if not entirely at listbuilding stage. At least be consistent in your misguided arguments. I have provided you a detailed explanation about why exactly you are wrong a post above. Your sports comparison is not valid here, because typical sports are not games of attrition and consist of large number of separate plays (literally more than a hundred of resets in case of basketball, a bit less in volleyball and several in case of soccer, connected only via total score). Imagine that after each goal in soccer a loosing team has to play with one or two players less and rethink if there wouldn't exist a point when reclaiming a win wasn't humanly possible.

nou wrote:
Large number of dice equalizes single event outcomes - like guard blob shooting, but it doesn't do anything for chains of causality connected events that constitute a game of 40K. In other words - non-average rolls of large variability events weight a lot more in early turns and on actions taken early (they cut down game history branches that are closer to root), even if common perception of "deciding roll" is that it happens late in game.

This is technically true, but a 40k game is sufficiently large that even "early turns" is a large enough sample size for the bell curve to start to apply. For example, the "early turn" equivalent in X-Wing might consist of a single roll of 3 red dice vs. 3 green dice involving a key ship, while a single shooting or combat phase for a 40k army still involves buckets of dice.

But yes, the sequencing of streaks of luck do matter even if the final outcome of the dice is a perfect "fair die" distribution, that's why I acknowledge that luck is a factor. It's just much less of a factor than it is in other games.

Even if you consider things like single round of shooting of 5 quads of 5 tacticals, the outcome of such round cannot be reasonably approximated by averages, because single events that roll small number of dice have much higher variance of results than a single large squad of 25 tacticals shooting together, and you base your decisions about what 2nd, 3rd, 4th and 5th squad of marines does on the outcomes of previous events.

This, however, is wrong, for three reasons:

1) 40k is a game of rolling buckets of dice, each with a low probability of success. A bolter shot has an 11% chance of inflicting an unsaved wound on a MEQ target. A single 5-man tactical squad is unlikely to do much even with above-average dice, so the vast majority of the time you're effectively treating those five separate squads as a single large squad and focusing their fire to gradually wear down a target. At best maybe you finish it off a bit early and the 5th squad is free to act, but you're not getting a lot of extra options by avoiding overkill.

2) 40k is not just a game of exchanging dice and stat lines. Even in 8th edition, with GW relentlessly trying to kill off the importance of maneuvering and LOS, position still matters. You might be able to come up with some theoretical math about how the dice are working, but on the table you've probably committed those tactical squads to a position where they have one good target (since you know, on average, that even 25 tactical marines aren't doing a lot of damage and will need to focus fire). Even if you get the information advantage and free up your 5th tactical squad to shoot at a separate target instead of adding to overkill how often are you really getting to deliver those extra shots in a game-changing way? Compared to all the times where you find that one bolter marine is within range to throw a low-probability shot at a random vehicle and nothing is really gained?

3) 40k is a huge game with lots of units. Those five 5-man tactical squads and their target are still a small part of the game. Even exceptionally good or bad dice luck with those 25-50 dice won't swing the game either way. Again, consider the X-Wing example where one exchange of fire between two ships, using fewer than 10 dice total, decides the entire game. Nothing involving those tactical squads is anywhere near as significant, and you're rolling many more dice to do it. It's just incredibly unlikely that dice luck is so overwhelmingly skewed that it makes a comparable difference in 40k.

The resulting game of 40K will be an utter abomination and won't resemble typical match in the slightest.

Well yes, but I'm not sure what your point is. The resulting game where one player rolls nothing but successes for their entire opening shooting phase and wipes the other player off the table would require such obscene luck to accomplish with actual dice that we're talking about "what if I win the lottery every week for a year" scenarios. It just doesn't have anything to do with reality.


Automatically Appended Next Post:

nou wrote:
I would like to remind you here, that you are the one who constantly argues how for at least 3 editions, games of 40K are definitely decided at most at the "who goes first" roll if not entirely at listbuilding stage. At least be consistent in your misguided arguments. I have provided you a detailed explanation about why exactly you are wrong a post above.

And I've explained why you are wrong.

And yes, I have argued that games of 40k are decided in list building, but that's an unrelated point. Choosing to optimize for winning while your opponent brings a random battleforce-style list is not a dice event, it's a difference in mindset towards the game. I have no idea why you'd think that's inconsistent with arguing that 40k is less dice-dependent than other games.

Your sports comparison is not valid here, because typical sports are not games of attrition and consist of large number of separate plays (literally more than a hundred of resets in case of basketball, a bit less in volleyball and several in case of soccer, connected only via total score). Imagine that after each goal in soccer a loosing team has to play with one or two players less and rethink if there wouldn't exist a point when reclaiming a win wasn't humanly possible.

You're missing the point entirely. In your hypothetical soccer game there would be a point where winning is effectively impossible, but getting there is a trend line of cumulative events, not a single discrete point. You don't go from 11 players to 1 because of a single discrete event, you get there because the other team scored several goals and you failed to defend. And scoring each goal is not a single discrete event. Yes, only one kick goes in the net, but if your team is in a good attacking position you're on a trend line of a favorable offense. If a defender steps in to cut off one particular shot maybe the attacking player holds their shot, steps to the side, and shoots from a slightly different position. Or maybe the shot is blocked and the attacker, being in a good position, recovers the ball and resumes the attack. Or lets say the attacker scores, and now it's 9 vs. 11. That's skewed, but not hopeless. Now the team with 9 players takes several shots at the opposing goal in an attempt to equalize the score and player count, missing all of them. Which of those misses decided the game? Etc. Singling out a particular event as "the one" is great for human psychology, but it doesn't model reality very well.

 Dashofpepper wrote:

nou wrote:

 Peregrine wrote:

1) Dice are IMO much less of a factor than in other games because of how many you roll. Over the length of a full game you're going to get close to that bell curve distribution for both players, and the game is designed around rolling buckets of dice with each die having a low individual probability of success. It's not like, say, X-Wing where you only have 3-4 units on each side and a single roll of 3 dice can decide if you win or lose.

This is probably the most often perpetutated myth on dakka and shows total lack of understanding how probabilities and statistics interact with eachother in games...

Large number of dice equalizes single event outcomes - like guard blob shooting, but it doesn't do anything for chains of causality connected events that constitute a game of 40K. In other words - non-average rolls of large variability events weight a lot more in early turns and on actions taken early (they cut down game history branches that are closer to root), even if common perception of "deciding roll" is that it happens late in game. Even if you consider things like single round of shooting of 5 quads of 5 tacticals, the outcome of such round cannot be reasonably approximated by averages, because single events that roll small number of dice have much higher variance of results than a single large squad of 25 tacticals shooting together, and you base your decisions about what 2nd, 3rd, 4th and 5th squad of marines does on the outcomes of previous events. This is really probability and statistics 101 and the reason why low discrepancy weapons/units are better competetive choices than theoretically powerfull units with huge swings in the damage output. You can however judge the overall performance of a given codex unit via "theoryhammer", because that is reasoning that is causal independet and will even itself over a course of many whole games.

And what Daedalus wrote is true - in every game of 40K ever played there is a point at which the winner is already decided, even if players are not aware of this fact yet, because every game, even luck-based, can be analysed by decision tree, and within each tree there is a point, in which remaining decision branches are too short to change the outcome. his may be the "who goes first" roll in case of badly biased matchup or it can be a single close roll in the last turn of the last round. The overall good casual game design targets deciding rolls to be as close to the end of the game as possible, but in sandbox games (and 40K is a sandbox game) it is literally impossible to achieve this, so in result most 40K games are decided rather earlier than later into a match (and no, X-Wing considered as a whole system, including all unusable cards and ships is no better than 40K in that regard, the main reason why it is perceived as ballanced better is that x-wing is much closer to CCG in terms of emotional/time involvement into collecting/changing lists, so it is far easier for players to chase the meta, so in turn it is much less of a sandbox game). This is also why many games introduce tiered win classes, because even after this deciding point in a game, there are still decisions remaning that can alter the exact discrepancy of players scores even if they cannot change the winner.


Automatically Appended Next Post:
And a little thought experiment to chew on - the next time you play, record the total number of dice rolled throughout the game (or you can even record actual results), and the next time you play, allow both players to utilise a pool of dice results - either experimentally collected or prepared based on simple division of entire pool into six categories of results. Such pool will, by definition, follow average distribution exactly. The resulting game of 40K will be an utter abomination and won't resemble typical match in the slightest.

You are correct - but we can't apply data points into a model to test accuracy without an algorithm to test, which in turn require variables, which brings us back to the root question of "How can we express the variables that make up the competitiveness of a given situation in a logical form?"

Take a stab at it!

The problem with your original question is that it's heavily dependent on large number of variables not included in original equation. Firstly, the exact tournament format - you need completely different formula for ITC, original Eternal War, CA18 or Maelstrom, because while you can "solve" ITC or Eternal War trees (by solving I mean cutting whole lot of branches during the pre-game stage of listbuilding and objective choice/placement and in ITC pretty much preparing whole detailed game plan before you reach the table), you cannot do this as easy or as definitely with CA18 and you can only prepare for probabilities of card draws in Maelstrom (this is why I personally find Maelstrom most interesting challange). Value of tactical prowess also changes greatly dependng on whether you play Eternal War on planet bowling ball, or a detailed objective based game on a dense, Necromunda like table. In this first case, the whole of so called "tacticall prowess" is target priority and positioning minutiae, while on a dense table it also contains planning entire rounds of movement and possible interactions along the way in advance, greatly increasing the complexity of decision trees. Another thing is that codex power discrepancies and themes in a "bowling ball Eternal War" scenarios can be so great, that your proposed equation does not function at all - there are matchups possible, that you cannot win even with maxed 0.85 total of A+C.

To conclude, I don't think that you can approximate complexity of 40K with such a simple equation. But, based on auticus' experience, I think that many 40K players would actually like that your equation was true and that they consider listbuilding skill as a part of tactical prowess parameter - hence the whole heated ITC vs CA18 debate we had here a while ago.


Automatically Appended Next Post:
@Peregrine: I'm not discussing with you, I'm lecturing you. I'm far beyond treating you as equal partner in pretty much any topic that gets discussed on Dakka.

nou wrote:
@Peregrine: I'm not discussing with you, I'm lecturing you. I'm far beyond treating you as equal partner in pretty much any topic that gets discussed on Dakka.

That's ok, I feel the same way about you and your blatant mistakes about probability.

(And in your opinions about "casual" games and game design, but that's a separate subject.)

I have to side with Nou's take on this. Even if you are comparing tactical squads, the first squad to shoot effectively not only does damage first, but also limits the damage that they will receive back. A minor hot streak early on could put a player down enough material to not come back.

And that ignores a bigger reality: that important units are often easier to kill, point for point, than chaff. Good (or bad) luck eliminating the opponents key units can effect strongly what the enemy can do for the rest of the game. The difference between leaving a knight with one wound left and destroying it utterly is tiny statistically, but massive in terms of game play.

 Polonius wrote:
I have to side with Nou's take on this. Even if you are comparing tactical squads, the first squad to shoot effectively not only does damage first, but also limits the damage that they will receive back. A minor hot streak early on could put a player down enough material to not come back.

Except, again, a minor hot streak can't do that nearly as much in 40k as in other games because even "small" events in 40k involve tons of dice. A round of shooting with those 25 tactical marines involves 25-50 shots, each of which has 1-3 dice. That's an entire game worth of X-Wing dice to resolve a minor exchange of fire between small units. A minor hot streak of an extra 3-5 dice simply can't do that much damage in this case, while a minor hot/cold streak of 3-5 dice in X-Wing can end the game in a single shot. And the kind of streaky dice luck that would make a meaningful difference in 40k would result in a 100-0 tabling in one turn in an X-Wing game.

And that ignores a bigger reality: that important units are often easier to kill, point for point, than chaff. Good (or bad) luck eliminating the opponents key units can effect strongly what the enemy can do for the rest of the game. The difference between leaving a knight with one wound left and destroying it utterly is tiny statistically, but massive in terms of game play.

But the point is that leaving a knight with one wound left isn't a single discrete event. If it's such a priority target that leave it alive for one more turn (though only hitting on 6s) is swinging the game you're going to keep shooting at it with everything you've got until it fails that last save. So yeah, it surviving in that case might be significant, but it's going to be rare that so many dice all went in favor of your opponent. Contrast this with killing a priority target in X-Wing, where everything might come down to a single roll of three defense dice and the difference between all evades or all blanks decides the game.

 Dashofpepper wrote:


I don't agree with that either. That presents armies in a vacuum, detached from the OTHER things that influence rankings. Consider this:

-In your example, 143 players took IK/BA/IG
-Let's say that the top 10 armies were all Imperial Knights.
-People therefore make the conclusion that Imperial Knights are the strongest army.

Welcome to Dakka ... I think you just nailed 90% of the feelings to current Meta



and this is not armies in a vacuum either .. that's a real world example from recent ITC event ... the top 50% of each of 5 rounds were IK+IG+BA or some variation .. because previous events had shown it to be a good list regardless player skill


the names you quoted added the skill / experience and took it to top 10.

 Peregrine wrote:

 daedalus wrote:
Depends on the army and point values I guess. Guard will smooth out much more than an elite army. I played 1000 points (Adepticon TT rules) of DA Primaris yesterday and probably threw less than 300 dice the whole game. Meanwhile, I've had 1000 points of guard that could probably throw that many dice in a round without breaking a sweat.

That's just proving my point. You played a small game (half the typical point level) with an elite army and you still rolled 300 dice. A game of X-Wing is typically decided by somewhere in the range of 10-20 dice.

Is it? Sidestepping the individual outcome vs whole game thing for the moment (even though specific outcomes are more important) is 300 dice is enough dice to expect even distribution? Dunno. Is it closer to even distribution than the teens of dice you roll in xwing? Sure. Does that mean anything? I am not sure.

I've not slept all night and this is approaching the "too much effort" cutoff, but I'll go anyway. There's probably a statistical model that can do it better than this, but I can't think and this is what I have right now:

I roll 300 dice and count the outcomes. Actually, I just had a computer do it for me. Repeatedly, even, so that it can build an average difference of outcomes between the most rolled and least rolled. The variance between the most rolled outcomes and the least rolled outcomes averages about 6.6% of the total dice rolled. Based on that, assuming you and your opponent are both rolling 300 dice each throughout the entire game, I would expect there to be a soft 13% ceiling to the amount of difference in performance 300 dice can get you, assuming one of you rolled ~20 more 1s than 6s and the other rolled the opposite. That's not insubstantial, but that's a pretty small game. Obviously that'll drop off as you increase the number of dice. It'd get down to about a 5% difference (give or take a percentage point) if you doubled the dice, which would be a 2000 point elite army game. Whether that's significant or not I'll leave up to the reader.

But that's a very broad look at it, because a game isn't a single event for which dice are rolled. Some action taken within the game is. And each outcome influences the next action you take, because events occur in linearity and your actions are finite and thus can be expressed in a tree like nou was saying. A single wildly swinging outcome on any given action is incredibly significant and much more likely to happen, particularly when you have an effective economy of ~15 actions per phase (average number of units at 2000 points based on scientifically glancing real quick at Dakka Army Lists) most of which will not be equally weighted in tactical impact (firing a captain's bolt pistol vs a LRBT) at that moment. Hell, you could collapse an entire branch of your tree simply by failing an above average number of armor saves on a pivotal unit, which becomes much more realistic of a worry than something as abstract as an entire game's worth of rolling bad. And you'll say that this is where the being a good player comes in to play, and I'd agree, but that still means that the importance of tactical skill is largely dependent upon the dice rolls. Generally speaking, I don't need to be as skillful if I roll above average outcomes as I would if I rolled below average outcomes, and thus it matters less.

And really, you may well roll an statistically perfect distribution on your dice across the game, and you may do it every time, but even if you break it down by turn and not by individual action, if the first couple turns of rolling are much lower on average than the second half of the game, you're probably gonna be pretty hosed regardless of all of this.

One thing I see here is that top player value skill higher than other players. I think part of that is that top players are already accounting things like codex, army building, and dice before the game begins.

I think the initial equation falls down (unless you are accounting for list building in tactical prowess, which I don't) because it lacks enough variables, I also think you are weighting tactical prowess too highly. Which I believe stems from the fact that you are assuming well built lists facing one another. Were I to try a simple equation (which I don't think really works) I would go with

T = Tactical acumen
L = list power
D = Dice
M = match-up - (even good lists have favorable, unfavorable matches)
F = Mission format

Then the result would be something like

x = 0.4L + 0.15D + 0.15M + 0.1F + 0.2T

I think you rate T so high either because you are including list building, or because you are only considering matches of top lists where things like list are largely nullified. A great player isn't winning with an awful list, but they never take those types of lists.


Automatically Appended Next Post:
One thing to note about dice probability over the course of a game is that it may well average out, but timeliness of those rolls matters as well in considering good dice, so if in say 300 dice you roll 100% average (50 of each result), but lets say those dice represent say Ork boyz in Combat in a given turn against plague marines, and when you are rolling you get all 150 of your 4s, 5s, and 6s rolling to hit. Then Roll and roll 150 1s, 2s, and 3s rolling to wound. This is 100% statistically average for rolling 300 dice, but the result will be you wounding 0 plague marines. Now that is ridiculously unlikely to actually happen, but it doesn't take much to keep your statistical average, and still roll poorly for the outcome.

 daedalus wrote:
is 300 dice is enough dice to expect even distribution?

Yes. Let's simplify the model and assume that those 300 dice each have a 50% chance of success. What are the odds that you roll at least 10% more successes than average? A whole 4.7%. That's less than a 5% chance to get even 10% more power through luck alone. Want 20% more successes from luck? That's a 0.03% chance. IOW, once every few tournaments you might get a small advantage in luck in a single game, and maybe once in your entire playing career you might have a dice advantage comparable to the advantage of taking guardsmen at 4ppm instead of the 5ppm most people think they should be.

And, again, that's for a half-size game with an elite army. Just playing at 2000 points and doubling the dice total to 600 brings the chance of getting 10% more successes down to 0.79%, and the chance of 20% more successes so low that the calculator I'm using can't even display such a small number (less than 0.0001%). Games where dice luck significantly diverges from the average are rare.

A single wildly swinging outcome on any given action is incredibly significant and much more likely to happen

Yes, but the point is that in 40k these events are much rarer than in other games. As you said, you have an average of ~15 units on the table, and even single-model character units are rolling at least 5-10 dice a turn. It's not like X-Wing where you have 3-4 units per player and a single roll of three dice can kill one of them and often effectively end the game, if you're in a position where you need to make a single roll to decide the game it's probably the result of a cumulative trend of choices and dice putting you into a bad situation and you're focusing too much on the final roll of the dice that made it official.

Hell, you could collapse an entire branch of your tree simply by failing an above average number of armor saves on a pivotal unit, which becomes much more realistic of a worry than something as abstract as an entire game's worth of rolling bad.

But again, these events are rare. Let's take something as basic as a tactical squad rolling armor saves. 10 models, 10 wounds to save against. That's an average of 3.33 dead, so what are the chances of rolling badly enough to lose an entire additional model over the average? Only 21%. To lose two models over the average it's down to 7.7%. And you have a 0.001% chance of failing all 10 saves and losing the whole squad. And that's for a ~150 point squad, out of a 2000 point army. The realistic contribution of dice you'll see is on the magnitude of plus or minus a dead tactical marine or two, hardly game-changing outcomes.

Now granted, if you define "pivotal unit" by point cost then a failed save kills more points worth of stuff and matters more but those units also tend to have stronger defenses in general and a much greater ability to survive wounds.

 daedalus wrote:

 Peregrine wrote:

 daedalus wrote:
Depends on the army and point values I guess. Guard will smooth out much more than an elite army. I played 1000 points (Adepticon TT rules) of DA Primaris yesterday and probably threw less than 300 dice the whole game. Meanwhile, I've had 1000 points of guard that could probably throw that many dice in a round without breaking a sweat.

That's just proving my point. You played a small game (half the typical point level) with an elite army and you still rolled 300 dice. A game of X-Wing is typically decided by somewhere in the range of 10-20 dice.

Is it? Sidestepping the individual outcome vs whole game thing for the moment (even though specific outcomes are more important) is 300 dice is enough dice to expect even distribution? Dunno. Is it closer to even distribution than the teens of dice you roll in xwing? Sure. Does that mean anything? I am not sure.

I've not slept all night and this is approaching the "too much effort" cutoff, but I'll go anyway. There's probably a statistical model that can do it better than this, but I can't think and this is what I have right now:

I roll 300 dice and count the outcomes. Actually, I just had a computer do it for me. Repeatedly, even, so that it can build an average difference of outcomes between the most rolled and least rolled. The variance between the most rolled outcomes and the least rolled outcomes averages about 6.6% of the total dice rolled. Based on that, assuming you and your opponent are both rolling 300 dice each throughout the entire game, I would expect there to be a soft 13% ceiling to the amount of difference in performance 300 dice can get you, assuming one of you rolled ~20 more 1s than 6s and the other rolled the opposite. That's not insubstantial, but that's a pretty small game. Obviously that'll drop off as you increase the number of dice. It'd get down to about a 5% difference (give or take a percentage point) if you doubled the dice, which would be a 2000 point elite army game. Whether that's significant or not I'll leave up to the reader.

But that's a very broad look at it, because a game isn't a single event for which dice are rolled. Some action taken within the game is. And each outcome influences the next action you take, because events occur in linearity and your actions are finite and thus can be expressed in a tree like nou was saying. A single wildly swinging outcome on any given action is incredibly significant and much more likely to happen, particularly when you have an effective economy of ~15 actions per phase (average number of units at 2000 points based on scientifically glancing real quick at Dakka Army Lists) most of which will not be equally weighted in tactical impact (firing a captain's bolt pistol vs a LRBT) at that moment. Hell, you could collapse an entire branch of your tree simply by failing an above average number of armor saves on a pivotal unit, which becomes much more realistic of a worry than something as abstract as an entire game's worth of rolling bad. And you'll say that this is where the being a good player comes in to play, and I'd agree, but that still means that the importance of tactical skill is largely dependent upon the dice rolls. Generally speaking, I don't need to be as skillful if I roll above average outcomes as I would if I rolled below average outcomes, and thus it matters less.

And really, you may well roll an statistically perfect distribution on your dice across the game, and you may do it every time, but even if you break it down by turn and not by individual action, if the first couple turns of rolling are much lower on average than the second half of the game, you're probably gonna be pretty hosed regardless of all of this.

Even sleep deprived you are right, perfectly understandable and coherent, but as expected, Peregrine understood nothing.

One thing I may add to this, is that in objective driven progressive scoring scenarios with enough LOS blocking terrain, 40K rounds commonly divide into independent pockets of interactions of single to few units against single to few units (Maelstrom is especially notorious in that regard) and 10 dice deciding events are very, very far from being "rare things to happen in 40K". 10 dice are nowhere near the volume that follows normal distribution. That previous decisions and dice results converged to such pivotal point? But of course. Could there exist an alternative solution and resolution of previous turns/rolls? Of course, that's the whole point of 40K being a game. Would there exist another such pivotal point in alternative game progression tree? Of course, as even from high-school level analysis course we do know, that if we start at the point of 50/50 chance of winning and arrive at 100/0 point of final game resolution, there has to be point of no return (or a divine intervention rearranging the tree), because the game is a linear, continuous path from one point to another on a decision tree. Were there solutions in which instead of pivotal shot there were pivotal turns of interchangeable shooting? Only in EW on sparce terrain there is enough redundancy in shooter/target selection for this condition to be true. Perhaps Peregrine cannot understand both of us, because he dismisses large chunks of BRB and ways people play this game as non-existing and has this abstract image of a single "archetypical 40K game" that everyone repeatedly plays out...

That said, it's possibly an argument for standardized competitive tournament armies.