Saturday, November 3, 2012

Winfluence

Along with the PBA making great strides in the live presentation of PSP Pro paintball this year they augmented that live presentation with player stats and a supporting website that is bringing a new level of insider info to fans and fellow paintball players.
The move to include stats as a way to both measure the pro players performances and create a way to talk about the players and the sport is a positive step forward.
But--there's room for some improvement and I want to start with the cleverly catchy 'winfluence' rating. At best it's a misleading stat as part of a player's ranking. On the other hand as an internal team assessment tool it's a potentially useful indicator.
Here's how it works: Winfluence is currently assigned a "weight" of 15% of the total player score. As a practical matter that means that the player with the highest calculated 'winfluence' receives a 15 and everyone else is some fraction thereof. Without debating the merits of the whole stats accumulation system (today) the larger issue is just what does 'winfluence' measure? It measures percentage of points won by individual players across a team. And as a result it rewards a wide divergence of ability on a single team. For example, the top 2 'winfluence' rated players play for the same team. The players win around 62% of the points they play in. Their team as a whole wins around 52% and the difference is what 'winfluence' is measuring. The next two highest 'winfluences' among players are also both on the same team but they win around 35% of the points they play in. In neither case is the rating one that measures actual success but only comparative success--and that comparison is only amongst teammates. As a consequence all 'winfluence' tells you is that team A or B has a wide variance of something (experience, talent, skills, etc.) on their roster. Conversely a number of players on the team that won the highest percentage of points played this past season have a minimal 'winfluence' rating or a zero. So as a practical matter all 'winfluence' did this past season is over-value some aspect of a player's play based on the deficiencies of his teammates. Removing 'winfluence' would reorder the player rankings dramatically.
Keep 'winfluence' as a comparative stat within each team but in terms of ranking players replace it with something more on point--maybe something like percentage of critical points played and/or won or lost. And by critical I mean decisive points; points where the team either wins, loses or could win or lose. Who performs in crunch time? Now that I want to know.

19 comments:

Anonymous said...

Winfluence shouldn't be considered alone; it's the balance of win rating and the two stats should be looked at together.

Win rating is the % of points the player starts that end up won. That stat alone would give an advantage to players on teams that win more, regardless of the player's own contribution.

By combining winfluence with win rating, you differentiate players beyond the performance of their team as a whole. If their team tends to lose a lot of points, but does better when that particular player is on the field, that should reflect in the stats.

Where PBA messed up is not allowing negative winfluence. IF a team wins a lot, but loses more when a particular player is on the field, that should be reflected in the stats as well. Normalizing the positive and negative winfluence over the 15% would also reduce the stat's impact amongst players with positive winfluence numbers by half.

There are other issues with the stats as well, like some categories being normalized and some not. For example the top G-rating player gets 30 points, matching the 30% claimed for Grating in the stats. But the top win rating player only gets 10.23 points, making that stat worth only 10.23% of the stats, not the advertised 15. This makes Gcount count even more than it's already out-sized impact.

Baca Loco said...

Anon
If you combine winfluence with win rating all you're still accomplishing is to moderate win rating with a value that can't translate to all players--unless the values are determined in a different way.
I also don't think negative winfluence offers any useful further refinement.

Ken said...

When do we get to watch Lil' Baller discuss his stats?

George Kalkowsky said...

+/- should be on there. It's raw, easy to understand, and shows talent, can't argue it.

George Kalkowsky said...

Could even be normalized. Your +/- divide by best +/- and result <=1. Multiply by 15, there's your win fluence. I'd argue I want a consistent baller more than I want someone who's only consistent in crunch time, +/- tells that story.

Anonymous said...

alprazolam price xanax yellow bar mg - xanax names

Anonymous said...

zolpidem ambien zolpidem 10 mg side effects - zolpidem 10mg cost

Anonymous said...

cheapest ativan withdrawal from ativan how long - died ativan overdose

Anonymous said...

diazepam without prescription diazepam for pain - does valium diazepam do

Anonymous said...

xanax online how much does generic xanax cost - where to buy xanax bars

Anonymous said...

buy xanax online forum xanax generic version - drug interactions with xanax xr

Anonymous said...

buy diazepam online no prescription diazepam half life - buy diazepam 10mg online usa

Anonymous said...

valium no rx valium 477 - buy cheap valium online no prescription

Anonymous said...

buy ambien online ambien cr overdose symptoms - ambien dosage guide

Anonymous said...

order carisoprodol online carisoprodol for opiate withdrawal - buy soma florida

Anonymous said...

soma online soma generic fedex no prescription - what class is soma drug

Anonymous said...

buy soma online carisoprodol metabolism - where can i buy soma bras

Anonymous said...

buy ambien online buy generic ambien online - many pills overdose ambien

Brittanybqll said...

buy ambien online ambien cr overdose symptoms - ambien dosage guide