Nba nerd

The NBA Draft, Poker, and Bad Bets

One of the things that interests me is that during the NBA draft, there are a lot of people expressing a lot of certainty about somehting that has a long history of unpredictability: how any one player in the draft will perform in the NBA. There are any number of people who are sure that Andrew Wiggins, Jabari Parker, Julius Randle, or [insert name here] will be good players in the NBA. A few google searches can find you similar stories in the draft previews of every draft. Remember how Memphis was so sure that OJ Mayo was going to be a star that they gave up Mike Miller (who had already proven himself to be good) and Kevin Love?

But...I like to know the odds!

Understand that I am not making an argument from hindsight. I am not interested in the actual player analysis here. My point is not that Memphis should have known that Love was better than Mayo -- it is that Memphis essentially gave up Mike Miller for the right to move up two spots in the draft, to pick Mayo (whom they wanted) instead of Love (whom they did not). To give up a player as good as Mike Miller for that privilege requires a degree of certainty that simply does not exist.

And that's the crux. I'm not writing to say "Haha, look how stupid Charlotte was for drafting Bismack Biyombo!" (Well, at least not for the reason you're thinking). I'm writing to say "Haha, look how stupid everybody is all the time for acting as if there were that much certainty involved here!" By now, all of us should have learned that there is no certainty around the future NBA performance of NCAA athletes.

Given this, you might think that there are no bad bets in the NBA draft. You might think that I disagree with how Arturo has graded this year's draft picks (aren't they all "incomplete"?). That if you're taking a long shot anyway, you cannot fault a team for taking Dante Exum ahead of Marcus Smart.

You would be wrong.

Imagine a no-limit poker game where your opponent just went all-in in a big pot for $1000 more. You have A ♠ K ♠ and the board is Q ♠ 8 ♠ 7 ♡. Let's stipulate that there is now $3000 in the pot. Should you call? Your first instinct might be to say "it depends on what he has," but you would, again, be wrong. Since there is no certainty here (unless he shows you his cards), you must assume he can have a range of possible hands, and determine your average equity against those hands. And since you have very good odds to make the winning hand by the river against any reasonable "range", given 3:1 odds, you should always call here (note that even if he shows his hand, and is holding the strongest possible hand (three queens), you have 25.56% equity).

Conversely, if you held A ♡ K ♡, you have no flush draw, which significanly lowers your equity against the range of hands he can have (even if he is bluffing sometimes), and you should always fold (unless you know exactly what he has, and that he's bluffing, perhaps because of the way he's licking his oreo cookie).

Furthermore, the results should not make you doubt your decision; with A ♠ K ♠, you are always correct to call, and with A ♡ K ♡, you are always correct to fold. Even though sometimes A ♡ K ♡ will win, and often A ♠ K ♠ will lose. This is because you are getting 3:1 odds, and the ace-king of spades will always have 25% equity, and the ace-king of hearts will have more like 15% equity. So, if you call with A ♠ K ♠, and lose, you should not berate yourself. And if you call with A ♡ K ♡ and win, you should certainly not congratulate yourself.

So...back to the draft. We know that many players in the draft bust, and that the vast majority will never be high-impact players. Even the widely acclaimed 1984 draft class contained only 7 all-stars, and 7 of the first round picks never played in the NBA. It's full of guys who were "ok" or worse (anyone remember Melvin Turpin or Lancaster Gordon?). From the long history of the draft, we know that draft picks are long shots; it's just hard to predict NBA performance based on NCAA (or Euroleague) performance. Michael Beasley and Kevin Durant had nearly identical college careers. Anyone who claims that he/she "knew" Durant would be a hit and "knew" that Beasley would be so bad that the eventually a team would pay him NOT to play is probably lying.

But we also know that not all longshot odds are equal. There is a difference between 2:1 and 3:1. There are some college statistics that correlate more highly than others to success in the NBA (primarily rebounding, steals, and shooting efficiency). So, there are good bets (players with high marks in rebounding, steals, and shooting efficiency) and there are bad bets (players with low marks in those categories), and there are bets that you should discount heavily because of extreme uncertainty (e.g. Dante Exum has never played against adults, and we have virtually no data on him).

Of course, that doesn't mean that the good bets are sure bets. Nor does it mean that the bad bets are sure to bust. Sometimes you call that all-in on the Q ♠ 8 ♠ 7 ♡ board with your T ♠ J ♡ and you hit the miracle 9 ♡. That doesn't mean it was a good call. Sometimes you call with A ♠ K ♠, and your opponent shows 5 ♠ 6 ♠ and he scoops a huge pot by hitting a pair of 5s on the river. That doesn't mean it was a bad call.

All the bets in the NBA draft are long shots. Lots of the players we think are good bets will bust, and some of the players we think are bad bets will go on to become all stars. But some bets are good bets, and some bets are bad.