Speaking at the MIT Sports Analytics Conference, Microsoft’s Bruno Aziza explained how “Analytics tell a great story.”
Unfortunately, analytics rarely get used by the mainstream sports media in this way. Instead, metrics often get cherry-picked to fit their story. As a result, the public never gets the entire unbiased view that analytics provide.
For example, the media focused attention on the Miami Heat’s dismal shooting percentage in crunch time, concluding that they couldn’t excel in the clutch. They did shoot one for their first 18 when tied or trailing by three points or less in final 30 seconds. However, most teams have a low success rate in these situations (because defense is tight, referees hesitate to call fouls, etc.). And there are many more clutch situations during games. A balanced analysis would have taken these shots into account as well.
Most importantly, all statistics need a sufficiently large sample before we can reach any conclusions. The Heat had taken just 18 shots which met the specific criteria used. That’s far too few to say that they were poor clutch shooters.
The same thing applies for the hitter who has gone 0-for-10 lifetime versus a certain pitcher. Should the manager decide that he can’t handle the pitcher? Of course not! It would take a sample of at least 50 plate appearances before knowing that for certain.
Presenting a player’s complete statistical profile requires looking at numerous metrics over a significant sample of games. That’s why it’s vital for agents to examine every piece of relevant statistical data possible.
In baseball, all encompassing categories like WAR make valuing players easier. But that is just the start. As explained last year in The Sports Resource Newsletter, role players can have greater impact in select circumstances. Sustained success in high leverage situations also adds to player value.
Yes, analytics tell a great story. But it requires thorough analysis to present the data in a way that builds maximum value for your players.