{"id":1321,"date":"2013-09-10T14:47:59","date_gmt":"2013-09-10T18:47:59","guid":{"rendered":"http:\/\/heavytopspin.com\/?p=1321"},"modified":"2013-09-10T14:47:59","modified_gmt":"2013-09-10T18:47:59","slug":"simpler-better-keys-to-the-match","status":"publish","type":"post","link":"https:\/\/www.tennisabstract.com\/blog\/2013\/09\/10\/simpler-better-keys-to-the-match\/","title":{"rendered":"Simpler, Better Keys to the Match"},"content":{"rendered":"

Italian translation at settesei.it<\/em><\/a><\/p>\n

If you watched the US Open or visited its website at any point in the last two weeks, you surely noticed the involvement of IBM. \u00a0Logos and banner ads were everywhere, and even usually-reliable<\/a> news sites<\/a> made a point of telling us about the company’s cutting-edge analytics.<\/p>\n

Particularly difficult to miss were the IBM “Keys to the Match,” three indicators per player per match. \u00a0The name and nature of the “keys” strongly imply some kind of predictive power: IBM refers to its tennis offerings as “predictive analytics<\/a>” and endlessly trumpets its database of 41 million data points.<\/p>\n

Yet, as Carl Bialik wrote for the Wall Street Journal, these analytics aren’t so predictive<\/a>.<\/p>\n

It’s common to find that the losing player met more “keys” than the winner did, as was the case<\/a> in the Djokovic<\/a>–Wawrinka<\/a> semifinal. \u00a0Even when the winner captured more keys, some of these indicators sound particularly irrelevant, such as “average less than 6.5 points per game serving<\/a>,” the one key that Rafael Nadal<\/a> failed to meet in yesterday’s victory.<\/p>\n

According to one IBM rep, their team is looking for “unusual<\/a>” statistics, and in that they succeeded. \u00a0But tennis is a simple game, and unless you drill down to components<\/a> and do insightful work that no one has ever done in tennis analytics, there are only a few stats that matter. \u00a0In their quest for the unusual, IBM’s team missed out on the predictive.<\/p>\n

IBM vs generic<\/strong><\/p>\n

IBM offered keys for 86 of the 127 men’s matches at the US Open this year. \u00a0In 20 of those matches, the loser met as many or more of the keys as the winner did. \u00a0On average, the winner of each match met 1.13 more IBM keys than the loser did.<\/p>\n

This is IBM’s best performance of the year so far. \u00a0At Wimbledon, winners averaged 1.02 more keys than losers, and in 24 matches, the loser met as many or more keys as the winner. \u00a0At Roland Garros, the numbers were 0.98 and 21, and at the Australian Open, the numbers were 1.08 and 21.<\/p>\n

Without some kind of reference point, it’s tough to know how good or bad these numbers are. \u00a0As Carl noted: “Maybe tennis is so difficult to analyze that these keys do better than anyone else could without IBM\u2019s reams of data and complex computer models.”<\/p>\n

It’s not that difficult. \u00a0In fact, IBM’s millions of data points and scores of “unusual” statistics are complicating what could be very simple.<\/p>\n

I tested some basic stats to discover whether there were more straightforward indicators that might outperform IBM’s. (Carl calls them “Sackmann Keys;” I’m going to call them “generic keys.”)\u00a0 It is remarkable just how easy it was to create a set of generic keys that matched, or even slightly outperformed, IBM’s numbers.<\/p>\n

Unsurprisingly, two of the most effective stats are winning percentage on first serves, and winning percentage on second serves. \u00a0As I’ll discuss in future posts, these stats–and others–show surprising discontinuities. \u00a0That is to say, there is a clear level at which another percentage point or two makes a huge difference in a player’s chances of winning a match. \u00a0These measurements are tailor-made for keys.<\/p>\n

For a third key, I tried first-serve percentage. \u00a0It doesn’t have nearly the same predictive power as the other two statistics, but it has the benefit of no clear correlation with them. \u00a0You can have a high first-serve percentage but a low rate of first-serve or second-serve points won, and vice versa. \u00a0And contrary to some received wisdom, there does not seem to be some high level of first-serve percentage where more first serves is a bad thing. \u00a0It’s not linear, but he more first serves you put in the box, the better your odds of winning.<\/p>\n

Put it all together, and we have three generic keys:<\/p>\n