I was on a winning streak for a while but then I stumbled onto quite a few games where my luck was really really bad.  I don't feel any of my recent losses were due ENTIRELY to luck because I have found mistakes in my play in all of them, but it just got me to thinking that maybe my winning streak was actually just a luck streak.  I've started to play on here a lot and the longer I play the more I wonder how much of winning is skill and how much is dice.

I've noticed that whoever runs this site has a lot of neat graphs of things like board usage and site registeration.  There's also global ratings, separate board ratings, and H rating.  All this gives me the impression that the dev behind wargear collects a lot of data and organizes some of it for us.  Has there been any effort to calculate the correlation between luck and victory?  (Specifically for the very highly rated players in 1v1 matches.)  And has the idea ever been mentioned to calculate a new rating similar to H rating but adjusted for luck on a per game basis?  Lastly, is there a way to glance at the luck in my finished games without opening them up one at a time?

Having played more RISK, cards, and fantasy sports than the average person over the last half a lifetime, I can definitively say that over the long run, the most skilled players win more than luck would suggest. The least skilled players lose more than luck would suggest. And the vast majority of people in the middle of the bell curve have their fortunes dictated by the roll of the dice, hand dealt, or coverages blown than by their own actions.

Agreed with Cram.

Luck, and especially the correlation between luck and victory, have been discussed many times on this forum threads. I would also like a stat like the one you mentioned Minimal. It would have to be a "dice luck" adjusted stat as luck is also the initial setting you get, your starting seat, if you get good dice at the right moment, if you get a set with 3 cards when your opponent has to wait his 5th....

The combinations of factors that contribute to a correlation between luck and a win or a loss are far too complicated to attempt to force into some kind of equation, and I wonder that it even would be helpful to do make some that correlate wins and losses to individual cases

Consider something like dice luck for instance.  First of all (and this has been discussed ad nauseam in these forums), but luck stats need context..

http://www.wargear.net/wiki/doku.php?id=general:luck_stats#interpreting_luck_stats

But even if luck stats were improved upon, some boards favor early good luck more than others, and larger boards tend to have end games where garbage luck stats can accumulate while a player runs the board, etc.   I was just thinking that the amount of work and effort we have been putting into discussions related to this would be better spent on developing a decent AI player for the site.  There is a relationship between the topics.

So, was your winning streak a luck streak?  There are many ways to get lucky on this site - dice is the most obvious, but so is starting position (both initial territories, but also seat position), who your opponents are (get lucky with less-good players joining - or a bunch get booted), or an opponent makes a mistake that benefits you and so on. 

If you can critically analyze the games and determine how and where either luck or skill determined the game then you will become a more skilled player and that will win more games for you. 

>Has there been any effort to calculate the correlation between luck and victory?

I do wonder about this though.  A game like hex has zero luck influence.  Some maps, played with 2 players are probably nearly 100% determined by luck.  Is there some way at getting at this statistically?  It would be a fascinating stat to have on the board page.

I would also find this interesting as a map maker to see.  I think better maps would usually (but maybe not always) have higher influence of skill & less of luck.

>> Is there some way at getting at this statistically? 

I don't think so - maybe if you just reduced it to dice luck, but luck manifests itself in many ways -

"..over the long run, the most skilled players win more than luck would suggest"
Well duh but I want to know *how much* more. 

"as luck is also the initial setting you get, your starting seat, if you get good dice at the right moment.."
Yes but unless you can come up with some objective way to quantify those things, we are stuck with expected dice kills and losses vs the actual results.  Luckily I think with Risk that dice have by far more impact on the game than cards or starting position.

"but luck stats need context.."
On a game to game basis you definitely can't learn anything from blinding gazing at the luck statistics without ever seeing the game.  But if we could take thousands of games and compare luck to victory I think it would be reasonable to estimate from that how luck-based Risk is on the whole.

"A game like hex has zero luck influence."
Never played Hex but people say this about chess but I think chess is very luck-based.  There is theory on thousands of openings and if you are lucky enough to take a game with me into a position with which I'm unfamiliar you've gained a large advantage by chance.  Luck does indeed extend far beyond dice.

@minamal - your first three points contain valid arguments, but I would disagree with your last point.  Games like Chess - (and for this site Hex and Go-Geared 0-sided) are, with the exception of who moves first, entirely deterministic. In theory, players are totally in control of the impact of their own moves.  

A chess master does not let a game go down roads (s)he is not comfortable with.  At the highest of levels, the better chess player often manages to take the position out of the other player's comfort zone.  In other words, the advantage is not gained by chance as you maintain. If you want to claim that luck is a factor here, you need to start considering that the player who had the less sleep the night before was unlucky.

Ozyman wrote:

>Has there been any effort to calculate the correlation between luck and victory?

I do wonder about this though.  A game like hex has zero luck influence.  Some maps, played with 2 players are probably nearly 100% determined by luck.  Is there some way at getting at this statistically?  It would be a fascinating stat to have on the board page.

 

I would also find this interesting as a map maker to see.  I think better maps would usually (but maybe not always) have higher influence of skill & less of luck.

A map stat showing the % of wins coming from each seat in games with the same number of players.

Could be shown as an NxN matrix where N is the max number of players allowed on the map. X-axis could be the Nth seat. Y-axis could be Number of players in the game. Grid could be filled with win %'s. A final column could show the total number of N-player games, so you can get a feel for whether the % is a useful stat. (as in, the third seat wins 100% of 4 player games, but only 1 game has been played with 4 players is not helpful).

M57 wrote:

@minamal - your first three points contain valid arguments, but I would disagree with your last point.  Games like Chess - (and for this site Hex and Go-Geared 0-sided) are, with the exception of who moves first, entirely deterministic. In theory, players are totally in control of the impact of their own moves.  

A chess master does not let a game go down roads (s)he is not comfortable with.  At the highest of levels, the better chess player often manages to take the position out of the other player's comfort zone.  In other words, the advantage is not gained by chance as you maintain. If you want to claim that luck is a factor here, you need to start considering that the player who had the less sleep the night before was unlucky.

Likewise, in competitive chess, your opponent is typically a known entity. Heck, chess masters even change their game to exploit weakness in computer programs which are otherwise much better able to predict future outcomes than humans. They are able to do so because they prepare in advance, look at old games, know how you're most likely to open. Unless we're talking about a friendly game between Uncle Buck and his 5 year old nephew, Angus, whose moves are almost entirely random and accompanied by horse and sword slash noises, there is no more luck in a game of chess than there is in a boxing match.

Of course, if you leave the map War a 1-star review because there is no warning that Australia has an extra border, I would expect that losing on the opening move of a game of chess is common occurrence, and that most of your dice rolls are accompanied by airplane and gun fire noises.

"Vrooooomshhhhh! Rat-a-tat-a-tat-a-tat! Aha! Three sixes! I'm so much better at this than you!"

Sorry. That was mean, and passive aggressive. But I laughed and it made me feel better in a petty way, so I'm leaving it.

Ozyman wrote:

>Has there been any effort to calculate the correlation between luck and victory?

I do wonder about this though.  A game like hex has zero luck influence.  Some maps, played with 2 players are probably nearly 100% determined by luck.  Is there some way at getting at this statistically?  It would be a fascinating stat to have on the board page.

 

I would also find this interesting as a map maker to see.  I think better maps would usually (but maybe not always) have higher influence of skill & less of luck.

It would be complicated, but not impossible, to try to derive such a thing.  For most of the boards here, though, the sample size would be small enough that it would be random noise.

The stat that I would look for to shed light on this is the variation in ratings, versus a predicted variation in ratings given the players you have.  A skill based game should have a wider variation in ratings.  A luck based game should have a smaller variation.  This would work better if you had a better rating system.

You cannot control if your opponent picks a very obscure opening you've only played once or twice.
Likewise you don't have control over whether you or your opponent will make a terrible blunder. 
Even Fischer, arguably one of the best players ever, blundered his bishop in the first game of the 1972 world championship match.  Did Spassky have some kind of mind powers to make Fischer make such a critical error in an otherwise drawn position?  No, it was out of his control.  Spassky got lucky.  And the fact that Fischer went on to dominate for the rest of that match only proves this point.

Here's a blitz game I played yesterday. 
http://a.pomf.se/f8cba.png
I sacrificed a pawn for an attack that never quite materialized and I found myself a pawn down and totally lost, only playing for a draw.  Then suddenly my opponent played 34.. Kf6?? which leaves the knight hanging and loses on the spot.  He resigned a few seconds after making the move. We have a name for circumstances out of your hands: luck (also god if you're religious).

Cramchakle wrote

Likewise, in competitive chess, your opponent is typically a known entity. Heck, chess masters even change their game to exploit weakness in computer programs which are otherwise much better able to predict future outcomes than humans. They are able to do so because they prepare in advance, look at old games, know how you're most likely to open. Unless we're talking about a friendly game between Uncle Buck and his 5 year old nephew, Angus, whose moves are almost entirely random and accompanied by horse and sword slash noises, there is no more luck in a game of chess than there is in a boxing match.

The only time "your opponent is a known entity" is when you are a GM and your opponents are all famous.  I generally have no idea who I'm up against at tournaments and even if I did it wouldn't help much because I can't do an internet search for their name and get anything relevant. 

"computer programs [are] much better able to predict future outcomes than humans."
Wrong. 

"They are able to do so because they prepare in advance, look at old games, know how you're most likely to open."
Funny you mention this now because the Carlsen-Anand world championship just started a few days ago and in the second game Carlsen shocked the world by playing the Caro Kann.  I'm looking at a database now and he's played it 8 times out of over 1600 games.  In the press conference after the match Anand said the only reason he didn't push for the win was because he was caught so off guard by the opening choice and had prepared next to nothing for it.

But anyway I don't know why I bother explaining this to you as you obviously have no clue about chess.

Minimal wrote:

  anyway I don't know why I bother explaining this to you as you obviously have no clue about chess.

I think he is talking to you @Cram so of course I'm going to have to agree with him..

..and of course as a corollary, Carlsen-Anand sure must have been lucky to have opened with the Caro-Kann. ..and how unlucky was it for Fisher that at the end of his turn he found his bishop had moved in Ouija-like fashion to the wrong square?

You're both right: I know very little of chess, and am not afraid to admit it. I shouldn't have stuck with the chess analogies, as my competitive play extends no further than the other nerds at the game shop. We're better suited to MtG battles. But I liked the story about Angus that popped into my head, so I stuck with it. I've got a nephew who calls it skill any time he wins at any game, and luck any time he loses. I was getting that vibe in this thread. (Cram admits a lack of expertise? That's a backpeddling.)

Anyway, chess wasn't the point. I think it would be a very broad definition of "chance" that includes other people's actions within a set of known rules - regardless of the game. That was the point. My claim is, of course, no more provable than the opposite stance. I'll be bold, though, and say that choosing a less optimal reaction to an opponent's action (no matter how unusual) and calling it luck is to disregard the spirit of the word.

As to the explanation of the larger chess world. Thanks. It provided some fun wikipedia reads, and I'm better for it. I sometimes come across articles claiming GM so-and-so beat whats-his-name or beat some fancy computer program for the reasons I stated above. The perspective on chess tournaments I shared is the one provided to the layman with a keyhole view by whoever it is writing articles about chess.

>(Cram admits a lack of expertise? That's a backpeddling.)

clever! 

"I've got a nephew who calls it skill any time he wins at any game, and luck any time he loses.  I was getting that vibe in this thread."

Okay but my OP asks the question of whether my winning streak was because of luck.

The reason I call his opening choice luck is because Anand is an e4 player with lots of experience with the Caro Kann.  That means Carlsen probably used time he could have been studying on the openings he's already good to prepare some lines in the Caro Kann instead, betting that Anand would not expect it.  It was a gamble.  If Anand even had a hunch that Carlsen would try something like that, he probably would have made sure to brush up on a wide variety of lines just in case.  Judging by the way the game went, he probably decided to just study Carlsen's typical e4 responses, ..c5 and ..e5 .  I could be wrong.  Maybe Carlsen is so confident that he will try a fresh opening every single game, but I highly doubt it.  I'll let you know as more games are played.

Below is an excerpt from a thread where I calculated my opinioned effect of luck on about 700 1v1 Wargear Warfare games.  It is for dice luck only and does not attempt to account for starting position or card luck.

http://www.wargear.net/forum/showthread/2173/Luck_Cumulative_Chart_Wargear_Warfare_1_v_1_games_SquintGnome

=============================================================

Here is my latest update for my luck trend for 1 v 1 Wargear Warfare games.  From my last posting I dipped down to about -150 and have now bumped back up to about -100.  As mentiond in a previous post, I track luck % which is luck/expected kills.  This is about .20% now.

Here are some other interesting stats:

Overall Win %     59%

Seat 1 win %     72%

Seat 2 win %     39%

Win % if game luck is positive     85%

Win % if game luck is negative    33%

Win % for seat 1 and pos luck     93%

Win % for seat 2 and neg luck    19%

I think this illustrates the strong impact of luck for Wargear Warfare 1 v 1.  Although all games will be impacted by dice luck, 1 v 1 games are more sensitive to it.  In addition, I think this board is also more sensitive than others to a lucky turn because you can tip the balance of a game in just one turn.  For example, if you win 3 territories on the first round. 

I am not offering this data for any other reason than that it is interesting - thought others might agree. 

Don't know what you're talking about Min, Chess is entirely a game of skill. There is no good or bad luck...only good or bad moves. Unlike Cram, I was upon a time a very advanced chess player, but I still find his analogies pertinent.

I think your reference to Fischer's blunder being "good luck" is a broadening of the term. I believe most people that complain about bad luck here (like me) are referring to bad dice and/or bad placement, both of which are not decision-making for you or your opponent(s). When I make a bad decision, I have only myself to blame. When I roll 14-44 and just miss eliminating a player, then said player cashes a set and goes on an elimination spree to win the game, I call it...horseshit...and bad luck.