I just got out of a game that I won, but had severely lower luck that my opponent.  What is the lowest luck that you have had and still won the game.  On this one I was at -7

http://www.wargear.net/games/view/82605

Beastlymaster

I don't have access to luck stats at the moment, but I remember that this one was pretty low : http://www.wargear.net/games/view/53269

Can anyone tell me what were my stats on this map?

-68, not bad :-)

"Lowest" stat is pretty meaningless when you compare them from different games.  -7 can be a worse luck stat than -68.

Would you rather invest $10,000 in a $100 stock and watch the stock price lose $68/share, or invest $10,000 in a $10 stock and watch that stock price lose $7/share?

So the luck stat has no "absolute" meaning but is relative? I thought it was comparing vs. expected values of kills/deaths - so wouldn't that give a basis of comparison across games?

Both are correct, in a game with say 100 3v2 attack rolls, there is a range of say -100 to 100 whereas with 1000 3v2 attack rolls, similar -1000 to 1000.  These distributions are approximately normal, but the variance is larger with more rolls.

Said in another way.  A luck of -7 in a game with 100 rolls is much worse than in a game with 1000.  The correct way to "normalize" these to make them comparable across all games is not known at the moment.  We have discussed it as a community and Mongrel and I just haven't had time to get out hands dirty and figure it out.

nice find for the avatar Alpha!

Thanks, Mongrel found it as he crossed the country to visit me last week and snapped a quick picture as he drove by.  I cannot remember where in the west it was, but I am pretty it was in South Dakota.

AttilaTheHun wrote:

So the luck stat has no "absolute" meaning but is relative? I thought it was comparing vs. expected values of kills/deaths - so wouldn't that give a basis of comparison across games?

Alpha says you are correct with your above statement, but the answer to your question is "no".

I'm just dredging this from memory so there's a chance I'm getting it wrong, but I once suggested that you can get a rough comparison from game to game if you divide your "luck stat" by the square root of the number of rolls taken.  For instance  your game is comparable to an imaginary game where a player ends up with a -10 luck stat after 70 rolls..

-10 / sqrt 70 rolls = - 1.2

-68 / sqrt 3376 (the number of rolls you had in your game) = - 1.2

I would agree with Alpha when he says that this system is flawed, and we all would be much more in favor of a number that is based on Standard Deviation, but the number of calculations required to do this become massive as the data piles up.  I have suggested a back-door method to do this, but I'm not enough of a mathematician to know how to execute it.

As things stand, my sense is that the LS / SQRT (Rolls) method is not too far off the mark once you get over 30 or 40 rolls, and it actually does have a relationship with how SD is calculated.

In 2 turns of a 1v1 ant farm I managed to go to 37.76 luck differential between my negative and opponent positive. Oddly I had significant bad luck all night, so hope the system will get it out of its system and start fresh tomorrow!

Perhaps an easy way to 'normalize' the luck stat to some extent would be to add and extra calc called Luck %, which would be [(Deviation from expected / expected) + 1 ].  So a +7 luck based on expected 100 (kills) would be 7/100 + 1 = +7% luck.  Meaning you killed 7% more than the probability expectation.  +7 on expected 1000 would be only +.7% luck, etc.  So it would factor in the number of rolls to add some extra meaning. 

The problem is integrating the weighting and timing of a poor (or extremely good) attack...a lucky attack that barely procures a hard-to-get continent bonus (or breaks one up) can have extermely important consequences on the gameplay results and might be a blip in the luck factor ocean. 

Thingol wrote:

The problem is integrating the weighting and timing of a poor (or extremely good) attack.

There are many, many ...many more aspects of analyzing luck.  But I would argue that this isn't a "problem", as you suggest.

WG Luck Stats are meant to be nothing more than a pure mathematical analysis of the dice thrown.  We in turn put that information in context when we say something like, "Of course I lost, I only had positive luck stats towards the end of the game when I only had four countries left".

M57 wrote:

Thingol wrote:

The problem is integrating the weighting and timing of a poor (or extremely good) attack.

There are many, many ...many more aspects of analyzing luck.

Yeah, I think calculating the complete true 'luck' of a game is roughly equivalent to writing an AI player.

We in turn put that information in context when we say something like, "Of course I lost, I only had positive luck stats towards the end of the game when I only had four countries left".

or the inverse (converse?), "Of course I lost, I had horrible luck in the end game when it really mattered."

Alpha wrote:

Thanks, Mongrel found it as he crossed the country to visit me last week and snapped a quick picture as he drove by.  I cannot remember where in the west it was, but I am pretty it was in South Dakota.

Correct; I was bored on I-90.

I played a game in which my luck score was -180.21.  I won the game and the 2 players that finished 2nd and 3rd had luck scores of 99.11 and 70.14  6 Players.

http://www.wargear.net/games/view/29442

I checked 20 games and noted the following:

8 players had the highest luck score and won

4 players had the worst luck score and won

8 players had the worst luck score and finished 2nd

The number of players were:

2 games of 4

14 games of 6

3 games of 8

1 game of 12 

and so my conclusion is:

Luck scores don't really tell us much as they are currently calculated.

RECON

RECON wrote:

I played a game in which my luck score was -180.21.  I won the game and the 2 players that finished 2nd and 3rd had luck scores of 99.11 and 70.14  6 Players.

http://www.wargear.net/games/view/29442

I checked 20 games and noted the following:

8 players had the highest luck score and won

4 players had the worst luck score and won

8 players had the worst luck score and finished 2nd

The number of players were:

2 games of 4

14 games of 6

3 games of 8

1 game of 12 

and so my conclusion is:

Luck scores don't really tell us much as they are currently calculated.

RECON

 

 

Alternative conclusion: Luck isn't all that matters in a game.

Agreed, luck is not all that matters.

Another important aspect of how luck affects a game is when you get lucky.  Just a few lucky rolls in the beginning of a game ,  say +4 on one turn that is enough to capture and hold a continent, may provide enough bonus units throughout the course of the game to allow you to sustain a continuing trend of negative luck and still win the game.

Alternatively, if you get into a 3-way tie in endgame, it might not have mattered how lucky you were to get there, but luck when you are trying to eliminate could be critical.

I play with a regular group of friends, and if one person starts getting too far ahead there is a good chance they will get ganged up on.  If the game isn't foggy, this can negates any big runs of good dice luck - bad luck can still affect you.

In the game in which my luck score was -180.21 (that is a negative number, just so no one is mistaken).  There were 19,464 turns.  My luck started going bad at 6,710  (-.50) and pretthy much kept going bad for the rest of the game.

I do believe when you are lucky is important, sooner than later, or you are out of the game.

I would rather have more of it than less.

However, even when I looked at 50 of my games it became apparent to me that I have probably over estimated the value of luck in games.

My biggest change was  in accepting the dice rolls of this site vs others and realizing I would just have to account for that in my strategy.

Based on what several of you have said to accurately represent "luck statistics" is more difficult than would be worth the time. 

Perhaps what we have is good enough to assuage the pain I feel when I have what I consider to be totally unfair f**^#ing, s@%^*y etc dice.  AND then I go check the luck stats and find out I was doing better than everyone else!!  Oh well, that just removes one more really good excuse for my less than inspired play. ha ha

RECON