I was reading an article the other day about how IBM was mining player statistics to come up with their <a href="http://www-03.ibm.com/press/uk/en/pressrelease/34864.wss">keys to a match-winning performance</a>. It does not appear to work all the time. Check out this picture.
[caption id="attachment_1519" align="aligncenter" width="300" caption="IBM Stats - Fail!"]<a href="http://porcupyn.files.wordpress.com/2011/06/ova_battle.jpg"><img class="size-medium wp-image-1519" title="ova_battle" src="http://porcupyn.files.wordpress.com/2011/06/ova_battle.jpg?w=300" alt="Kvitova vs. Pironkova" width="300" height="200" /></a>[/caption]
Focus on these keys to the match for Kvitova:
- win fewer than 3% of points at net
Here is where blindly following some weird lifelong stats for a player and dumping it into their Wimbledon coverage gets IBM in trouble. Think for a minute, why would you want to intentionally win fewer of your points at the net? Makes sense if you are a baseliner, makes sense if you are on clay or hard courts, but if you are on grass, surely this stat should be out the window!
- win fewer than 8% of points won with a winner
But wait, now it gets more harebrained. Winning fewer than 8% of points with a winner? Maybe she is historically the Rahul Dravid* of Tennis, i.e., merely sits and waits for her opponent to commit a mistake. Even then, fewer than 8% of points with winners? Sounds unreal any day! However, again, this is Wimbledon, where one can hit winners much more easily than at, say, the French Open where, for instance, the ball will tend to sit up on drop shots.
* = this is not meant to insult or slight Rahul Dravid, who is the Indian cricket team's rock of Gibralter!
[caption id="attachment_1519" align="aligncenter" width="300" caption="IBM Stats - Fail!"]<a href="http://porcupyn.files.wordpress.com/2011/06/ova_battle.jpg"><img class="size-medium wp-image-1519" title="ova_battle" src="http://porcupyn.files.wordpress.com/2011/06/ova_battle.jpg?w=300" alt="Kvitova vs. Pironkova" width="300" height="200" /></a>[/caption]
Focus on these keys to the match for Kvitova:
- win fewer than 3% of points at net
Here is where blindly following some weird lifelong stats for a player and dumping it into their Wimbledon coverage gets IBM in trouble. Think for a minute, why would you want to intentionally win fewer of your points at the net? Makes sense if you are a baseliner, makes sense if you are on clay or hard courts, but if you are on grass, surely this stat should be out the window!
- win fewer than 8% of points won with a winner
But wait, now it gets more harebrained. Winning fewer than 8% of points with a winner? Maybe she is historically the Rahul Dravid* of Tennis, i.e., merely sits and waits for her opponent to commit a mistake. Even then, fewer than 8% of points with winners? Sounds unreal any day! However, again, this is Wimbledon, where one can hit winners much more easily than at, say, the French Open where, for instance, the ball will tend to sit up on drop shots.
* = this is not meant to insult or slight Rahul Dravid, who is the Indian cricket team's rock of Gibralter!
No comments:
Post a Comment