Interactive Tableau Chart: https://public.tableau.com/profile/chad.gordon#!/vizhome/Step12-PointWinProbability/Step12-PWP
(Quick note: all PWP in the chart are for Stanford. Therefore if PWP hits 100%, Stanford won the point. If PWP hits 0%, Wisconsin won the point. Anything above 50% means Stanford has the advantage, anything below…you get the idea. I find it much cleaner to use a single line rather than dueling lines that simply mirror one another for each team.)
So…we’ve talked about Point Win Probability (PWP) in recent posts. Simply translated: how likely are you to win the point, given the situation. Perfect pass, maybe now you’re at a 65% chance to win the rally. Just kidding, you overpassed the serve, you’re in some trouble and you drop to 25% chance of winning the rally.
Using this method, we can look at any and all contacts during a match. We then turn around and present this is a visual format so that coaches and overzealous parents no longer have to rely on feel, but rather, data. Much like NFL or NBA or NCAA win probability charts you may have come across, we simply look at how each unique state alters the likelihood of winning. While many publicly shared charts highlight trends over the course of a game – we offer a point by point timeline to start.
In the chart linked above, we look at the 2019 NCAA Championship between Stanford and Wisconsin. In subsequent posts, I’ll dive deeper into how to use this information – since at the end of the day, if we can’t do anything with the numbers…they’re pretty useless. Things like: Expected Value Added (eV+) and Expected Value Over Expectation (eVOE) taken at a player level, team level, situation level – can all add valuable insights in who is performing well, where to focus attention in practice, and how different playing styles materialize statistically.
This is exciting because now that we have all contacts converted into the universal language of Expected Value (eV) / Point Win Probability (PWP) it unlocks a flurry of other things that we can turn our attention to.
Just a quick note about where the values come from in the above chart. We basically look at how well Stanford has done historically (against all teams) throughout 2019 given all of the different situations – then we look at how well all teams have done against Wisconsin in those same situations and average the two. For example: Stanford might have a PWP of 70% on a perfect pass, but teams playing against Wisconsin this year have only managed a 60% PWP on a perfect pass. Therefore, we set the Stanford vs. Wisconsin specific perfect pass PWP at 65% in this hypothetical. The same applies to all other skills and touches.