Power Ratings I

Time for power ratings. The theory of the win/loss %s is that after a long period of time, everyone will play with a random group of players and all the teammates and opponents will be evenly distributed. Thus, your personal win % would be representative of how you compare to the other players. The problem is that it will take a long time before there is true parity in the teammates and opponents everyone plays with. Any given week or month you could have a very good or very bad run of players in your games (as well as good or bad runs of your own performance). The new answer is Power Ratings.

With PR (Power Ratings) I actually analyze all the games you have played and look at who your teammates and opponents were to see how they compared to the average. If you play with guys who usually win, you should win more often than not. On the other hand, if you face tough teams more often than average, that would negatively impact your chances of winning.

So, here is my first PR rankings.

The columns are:

1. Whom

2. Games

3. Win % (Actual)

4. Average Team (the average of the sum of the win % of the 3 teammates you played with... 150% is overall average)

5. Average Opponent (the average of the sum of the win % of the 4 opponents you played against ... 200% is the overall average)

6. Schd Strg - Strength of Schedule - subtracting the opponents from your team (with 50% for you) - 0% means you have played with and against completely average teams overall. Negative means tougher schd, positive means easier.

7. Pred win% - This is what your win % would be with those players if you were an exactly average player

8. Power Rating - your win % divided by your predicted win % - 100% means you've performed exactly average.

This chart has all kinds of crunched numbers in it. I'm gonna show the chart sorted a couple of ways to highlight a couple of points. This first version (above) is sorted by difficulty of schedule. I have taken the sum of your 3 teammates win % in all your games and taken the sum of your 4 opponent's win % in all the games also. If everyone was average, your 3 teammates would add up to 150% (50% each) and your 4 opponents would add up to 200%. So, if your opponents average more than 200%, you have faced teams harder than average. Also, if your teammates add up to less than 150%, you teammates have been worse than average. So, I have calculated the % of games you should have won (if you were exactly average) based on these numbers.

So, if you look at the chart, Tony's average team in all his games was only 146%, so if he were average, his team total would have been 196%. On the other hand, his opponents average team is a whopping 209% (that is huge, statistically speaking). So in the average game Tony has played, he has been 13% underdogs. As 13% underdogs, you would only expect them (if Tony were average) to win 39% of the time - based on the historical data. So, while Tony only has a 37% win percentage, that's not nearly as bad when you consider that with his teammates and against his opponents, he should have only won 39% anyway.

Other guys at the top of this chart who have been cursed by tough games are Ray, Sam, Shad & Gary. On the other side of the spectrum, are the guys who have had the biggest advantage in the games they've played. Good teammates and weak opponents. The runaway benefactor of being on good teams is Pride. Pride's 27 teams have averaged 172% (57.3% for each of 3 teammates). That's ridiculous (that's like having Steve, Nate and BK as your teammates - ON AVERAGE). He's only played a few days and played many games with James, BK, etc. His 4 opponents have averaged 199%. So, 172% + 50% for him if he were average, results in his teams being a 23% favorite (on average) in all his games. A 23% favorite wins 69% of the time at noon hoops. So, actually, Pride's 63% winning % isn't unexpected. Other guys with easy games on average have been James, Steve, Jack & Luke.

So, now that I know everyone's teammates and opponents, and I know how to translate team favorite percentages into predicted winnning percentages, I came up with the PR. It is simply your winning % divided by what your winning % should be based on your teammates and opponents. So, if you have won 54% and you should have won 54%, you are 54/54 or 100%. 100% theortically means you have performed exactly on average. Better than 100% means you've won more than you should have and less than 100% means you've won less.

That's a lot of words, if you have any questions, ask or email me.

So, here is the same chart sorted by Power Rating...

Of course, functionally what happened is that players with tough strength of schd are higher in the PR than they were when just looking at their win %. Conversely, guys who have played with better teams moved down the ladder in relation to their win %. I would predict that in the future, adding the power ratings together for a team will be a far greater predictor of the winners than just adding their winning percentages.

The bottom line is that 130% PR means that James and Anthony are personally winning 30% more of their games than the average player would with their teammates and opponents. Myself, at 88%, means that I win 12% less than an average noon hoops player would with my teammates and opponents so far. Crap.

I'll have today's stats later (these charts don't include today's stats).