What percentage is luck involved in a games outcome?

[axis_roll]

@Frood:

I agree, there probably are other factors besides skill and luck. However, the question seems to assume that those two are the only ones, so as the question is stated, the answer has to be 100%.

My point is that as stated, the question by definition means that skill is NOT a factor in deciding the game between the two players. Since it is equal, you cannot say that one or the other play won on the basis of skill. With skill eliminated, that leaves luck as the only factor contemplated in the question.

I already said the spirit of the question didn’t include skills being exactly equal.

I was trying to eliminate the Skill factor as a determinate of the outcome in an effort (as you so aptly put) to measure how much is luck (the only other ‘remaining’ factor in a game).

If I had asked
“how much luck is involved in A&A?”

The discussion would’ve been about players of differing skill levels, and then, of course, the discussion would’ve spit off into measuring skill… etc… blah blah Blah!

This is a “feeling” kind of question.

How about if I asked:
“How much do you feel luck is a factor in determing the outcome of A&A games?”

[newpaintbrush]

what if the players are equally skilled, but one player is more skilled at taking advantage of strategic openings that occur because of the results of a lucky attack, but the other player is more skilled at minimizing the problems that occur because of the results of an unlucky attack?

WAT THEN?

I think for players of equal skill, luck determines around 30% of the games played.

[frood]

If skill is not exactly equal (which only happens in theory, not in real life), then the importance of luck depends on the degree of difference in skill.

E.g. if I am playing someone who is just a little better than me, then I might beat them sometimes on account of luck. Luck will be a major factor in determining the games.

I cannot beat a superior player on skill alone. On the basis of skill, they have me beat.

If on the other hand I am playing someone who is WAY better than me, or someone who is way worse than me, then I will lose every single game, or win every single game. Luck becomes an insignificant factor when skill differences are large.

Now, since the significance of luck varies depending on the skill difference, you need to know exactly how big the skill difference is to determine how significant luck will be as a factor. “Roughly equal skill” does not give enough precision. Depending how roughly equal skill is, luck could range from a 95% factor to a 30% factor.

Do you see how the importance of luck varies according to the amount of difference in skill?

Look at it this way: going into a game, it is easier to predict the outcome when the skill difference is large. When there is little skill difference, it really is anyone’s guess as to who will win - you might as well toss a coin, because on the basis of skill you can’t predict. Thus luck is a bigger factor in that game.

[Guest]

I would argue that Dan.

There is a 3rd element, and one that I have found to be the PRIMARY determining factor in numerous recent games…
ERRORS.

SKILL is what a person has in terms of knowledge, ability, etc.  And based on my win/loss record, I think it is fair to say that I am one of the more skilled (in the top tier actually) of players on this site.

LUCK has come into play in a few recent games… the recent Tournament loss as a notable example (zero point 4 freakin percent probabilities…)

But an ERROR is different from skill.  And simple mistakes (for example, in my current game me missing a free liberation of Norway by Germany, or Jen’s strafe of Caucuses instead of the safe and easy route through Kazakh because she overlooked it) often times are the REAL factor in any given game.

Mistakes are NOT a lack of skill (unless you keep making the same mistakes over and over, never learning from them).  And I personally think that, skill levels being approximately even, that ERRORS will decide the game more than 2/3 of the time.

[frood]

I think really this boils down to semantics, because I have a different definition of skill that includes mistakes. I think the idea that a given player has a fixed amount of skill is a bit of a fiction. Rather, players will show different degrees of skill between games, and even within the course of a game.

The frequency with which you make mistakes, and the severity of those mistakes, is really a measure of your skill. Someone who screws up all the time is really not skilled, and someone who almost never does, is very skilled.

A mistake is basically an un-skillful move. In a way, most moves (except most of mine :) ) are mistakes, some just worse than others, because really, a mistake is a move that is not “perfect”. The severity of the mistake is just a matter of how far from perfect that move in.

So if I make a lot of mistakes in a game, I would say that I was not very skillful in that game. It will slightly decrease my overall “skill” measurement, which is some kind of average of how skillfully (ie. how mistake-free) I have played all my games. As I get older and senile, I will make more mistakes… I guess there’s skill in executing your moves, and a sort of meta-skill that indicates how skillful you are likely to be in a given game.

What you are describing I would call knowledge, experience, which will help you to play more skillfully. An error of oversight does not mean you are less experienced, but it does mean that that one move was executed with less than perfect skill.

But, it’s just semantics really. My beef was with the question, which is a matter of logic - “all else being equal” etc.

That’s what you do in a scientific experiment. You try to control all the other variables, so that you can measure the effect of one variable. Now, once you have controlled for all else, the one remaining variable will be the one that accounts for all of your results.

[Lucifer]

@newpaintbrush:

I think the best players are those that are able to exploit the openings left with good luck on an attack, and minimize the problems given with bad luck on an attack.

Those that play low-luck just can’t handle it!   :roll:

I disagree very much indeed, Sir!
But you actually got a point in your reasoning.
I prefer low luck, but I also play a lot of dice games.
In some games I played, multiplayer (dice) games, my observations concludes that this is sometimes a psychological factor.
If the attacking player uses overhelming force, but the dice gods are mad, then this sometimes leads to:

http://en.wikipedia.org/wiki/Acute_stress_disorder

and in severe cases this may develop into: http://en.wikipedia.org/wiki/Posttraumatic_stress_disorder

Seriously, do you want to fill up the psychiatric hospitals with A&A players diagnosed with combat fatigue?

I’m gonna write to health politicans in the entire western world, low luck shall be mandatory!!!

[axis_roll]

@Lucifer:

… low luck shall be mandatory!!!

In My Humble Opinion,
The game is totally different when playing low luck.
You have to think and play differently.

It’s a nice change of pace, but I would not want all games to be low luck.

There are no guarentees in War.

[Guest]

Personally, I think Low Luck should be relegated to the House Rules area, since it is a very significant change in the rules of the game from what is printed in the OM, via the Erata on AH’s web site, and from all versions of LHTR.

I think Nuno in another thread made a smart a$$ comment about custom rule sets that he felt were created as ways for players who can’t win using normal rules to be able to change things enough so that they CAN win.  If one grants that to be true, then Low Luck as a House Rule Variant to allow a player who otherwise cannot win to have a chance at winning would have to also fall into that category…

[Lucifer]

@ncscswitch:

I think Nuno in another thread made a smart a$$ comment about custom rule sets that he felt were created as ways for players who can’t win using normal rules to be able to change things enough so that they CAN win.  If one grants that to be true, then Low Luck as a House Rule Variant to allow a player who otherwise cannot win to have a chance at winning would have to also fall into that category…

Remember the classic boardgame? Played with original OOB rules.
I calculated that if UK+US used all ipc on research, then either US or UK would discover heavybombers rnd. 2,2 on average.
No TTL, 3 dicerolls, without “house rules” the game is unplayable if ppl use this flaw in the OOB rules.
And OOB rules both in revised and classic there’s no mention on bid……?
So bidding is a house rule…?
With this kind of thinking we’re all infidels…

[trihero]

Personally, I think Low Luck should be relegated to the House Rules area, since it is a very significant change in the rules of the game from what is printed in the OM, via the Erata on AH’s web site, and from all versions of LHTR.

I agree completely. Low Luck is a decent strategic modeling tool, but in practice it is indeed different in many ways.

The most obvious way in which it is different is in small battles. For instance, 2 fighters will always hit one infantry in LL. But in practice, they don’t quite hit one infantry on mathematical average, because even when both fighters “hit”, only one infantry dies. So 75% of the time 1 infantry dies, and 25% it doesn’t. So in full luck the “average” chance to hit the infantry in one round is really more like 75%, not 100%.

Low Luck is fairly accurate in larger battles when all of the attacking/defending units have a target to hit by themselves instead of “overshooting” on one target.

[Romulus]

I agree.
I think that if someone wants to reduce drastically the numbr of dice involved in a battle a more viable options is to use Combat Table (as in classic hex wargames).
Considering relative odds of the attacking and defending units combat power for defining the possible result and then choose randomly with a dice.
The result of the battle is almost totally defined by the relative strength, only the detailed result of the battle is random.
In such way it should be possible also to consider weather effect and terrain features. It is again an House rules, but differs from low luck, returning the planning in the field of strategy and tactics.

[Lucifer]

The problem I see with dice games is that too often in a game, the dice is really screwed.
Generally, skill is more important than luck, without tech!
Even if 2 playes seem to be just the same level, they are not able to play equally good, not in most games.
In a case if 2 players play each other, and they have played many games against each other, then they “know”
the opponent, and such game if often decided by luck.
I play only in the lobby (a.t.m.) and there is a BIG difference in skill and experience between different players.
Again, if your stats says you win way more than 50% you’r a good player.
Problem is “this certain game we play” is decided by major luck and minor skill……
I met some quite experienced players in the lobby, and I can’t understand why so many are
so reluctant to try low luck. For me it’s okay to play with dice sometimes, but if I could chose,
I would play 4 of 5 games with low luck setting.
For me it’s fun to win, because I win to little and lose to much, under 50% victory stat makes me unhappy.

So players who win a lot, who play reg. dice, you would not win if you play with low luck?

[Guest]

@Lucifer:

So players who win a lot, who play reg. dice, you would not win if you play with low luck?

I would have to learn and perfect a whole new set of parameters for my combat and defense.

[Romulus]

I usually play face2face games.
I may say that we do not feel that dices screws our combat.
I see situation in wich dices assist player more than expected but let me to say that often this happens when few dices are involved (particularly first turns).
We plan our attacks calculating die points, and in may personal experience, on average inflicted and recived hits differs not too much from the expected values.
Moreover players that scream for being screwed by dices are often lucky in following battles, so we have learnt to not complain too early, because often the dices give back what they taken before.

We are used to plan accounting risks, trying to have always a reserve plan.
With low luck A&A become more “arid”, with less risk and less excitement.
It is how you play poker with cards discovered, who ever will bet a cent if the one near have a greater hand?
Moreover where is the uncertainity of planning? Throwing only one dice is a marginal effect.

I am wondering about the randomization of dices results of the online games. Because for me the problem is not the randomness of real dice but the randomization of the pseudo-random generators used in on line dice rolling.
Maybe Frood may answer to this doubt.
Frood do you thing that pseudo-random number generation may have periodical flutuation? I mean oscillation of the average of the random generator regarding to the theoretical average of the dice (3.5).
Have someone ever tried to make a moving average analysis or some other kind of sampling to detect if there are periodic oscillation?

[frood]

Not quite sure what you’re talking about but I’ve tested my random number generator and it will average a roll of 3.5 on a d6.

[trihero]

Not quite sure what you’re talking about but I’ve tested my random number generator and it will average a roll of 3.5 on a d6.

That’s the thing with odds, in real life you can’t actually roll a 3.5. For instance if you’re fighting a tank, you have to kind of think of what happens when it hits, and when it doesn’t, because you can’t exactly go on the math which says it will kill half a unit on average, since you can’t kill half a unit (bb ships excluded lmao).

[DarthMaximus]

Just out of curiosity.

Rolling 100d6:
(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6)

[DarthMaximus]

I forgot about the ording function.  LOL!  :-D

At least it’ll be easy to count.

1’s:  14
2’s:  13
3’s:  21
4’s:  15
5’s:  22
6’s:  15

1-3:  48
4-6:  52

Little high on the 3 and 5’s, but close enough for me.

[trihero]

Lmao I too, at first, was like wtf  :-o :-o :-o you rolled it exactly in order on individual dice?

[Romulus]

@Frood:

Not quite sure what you’re talking about but I’ve tested my random number generator and it will average a roll of 3.5 on a d6.

I am not sure of being able to explain it better Frood! But I will try.

First I am not saying that dice rollers are not working well.
Really a dice roller, generating number from a pseudo random sequence, it is more “precise” than a real dice.
DarthMaximus showed with the test of 100 dices.
The problem I am trying to describe is related with the sequence of results, that is not possible to see from the preceding post because dice results are re-ordered, in ascending order.

I am trying to say that observing the results on a smaller period  (for example every 10 results) it could be possible to found that results are almost all above the average. While in the following period they are almost all under the average.
If you roll dices and get in the "above average " sub-period you has been screewed.
If you roll dices and get in the “under average” sub-period then you have been lucky soring more hits.

I am not saying that the dice rollers are not correct. The nature of pesudo casual sequences grant that the average is 3.5 but may arise problem with smaller sequence of the results.
I try to explain it with a simulated experiment. Consider the following sequences of results :
(a) 1 6 2 4 3 5 2 4 1 5 3 6 3 5 2 6 1 4: Average 3.5

(b) 4 6 5 4 6 6 5 5 4 3 1 2 3 1 2 1 2 3: Average 3.5

© 3 3 3 4 4 4 5 5 5 6 6 6 1 1 1 2 2 2 : Average 3.5

All the above sequences have the same average but going in combat with the last one © is not a beutiful thing! If I attack with 6 inf and 6 tanks I score 0 hits. Then defender defends with 6 infantry and score 6 hits!
The second sequence (b) has the same problem, having in the first half sequence (4 6 5 4 6 6 5 5 4 ) with an average of 5. The second half sequence is (3 1 2 3 1 2 1 2 3) with an average of 2.
The first sequence (a) suffers less from this problem, first sub sequence is (1 6 2 4 3 5 2 4 1)  averaging 3,11 the second subsequence is (5 3 6 3 5 2 6 1 4) averaging 3,88.

So my point is: pseudo random sequences have such kinds of oscillation? I mean on the long run they are almost “perfect dices” but analyzing the sequence in smaller segments has ever been made?
This is only an hypotesis, I may be wrong and this is not a real problem.

Addendum:

I made an experiment in the threads regarding dice roller, using 100 dices@1, that are not reoredered. Results are:
Considering sub sequences of 10 results, and calculating the number of values between 1 and 3 and the avereage of each sub sequence, we have:

Average  hits (1,2,3)
3,1 0
2,8 3
3,4 2
3,6 2
3,3 2
4 1
3,4 2
2,7 3
3,8 2
3,2 3

Each sequence of 10 numbers is “well equilibrated” (never more than 3 hits)
there are sequences like the sixth that averages 4 and get only one hit. But it seems that my hypothesis is not completely true.
There is an oscillation of the averages but its seems not a problem that may really hurt.