Wednesday, June 12, 2013

Detroit Petanque Club's Point System Updated & Corrected

After some careful examination, I found the old club spreadsheet to be lacking in some important data crunching features.

Here is the updated club's score table.
As you can see there are some interesting statistics going on here, namely Winning-Percentages and Point-Spread Averages.

The Winning Percentages show how likely it is for a particular player to win his or her game.

     Nick leads the pack at a 67% win percentage, followed by Abu (at 62%) and then Alpha (54%)

     Win Percentage =  (Wins x 100) / Total Games

But winning is only half the battle, you also want to measure how "well" you've won, in essence, the point spread average.  Winning 13/1 is a better point value than winning 13/12, likewise losing your game 1/13 is much worse than losing 12/13.  The point values (see below) are awarded for those spreads.

     Abu tends to win his games with a bigger point spread 2.04 (and his losses are more narrow margins).  He's followed by Me, Nick and Alpha (close averages 1.96; 1.95; 1.94 respectively)  Sam has a 1.93 but since he has only 3 games in the sampling, his average doesn't have enough data points (yet) to note it as statistically sound.

What does this mean?  Well, clearly, if you had a team composed of Abu, Nick & Alpha - then you'd have quite a powerhouse, statistically speaking.  It also means that players can think of themselves in terms of not only how often they win, but how well they win.  As for team composition, we can also try out forming teams that are statistically more evenly matched (of those that show for game play - obviously).

No comments: