Posts Tagged With: Statistics

Nassim Nicholas Taleb Called It

Nassim Nicholas Taleb saw the market downturn coming 4 years ago, and was prepared to profit from it. He is a font of pessimism about banks and the market, and one of the most fascinating men alive.

Taleb's argument is that statistics and probability can only do so much to quantify and manage risk. Statistics and probability are great for games like craps, where the events are well-defined and you can calculate the probability and the expectation of every bet. Hedge-fund types can apply these statistics to make small gains in everyday market trading. The problem is that the small gains are overwhelmed by sweeping disasters, bubbles or failures in the market (or in craps, the likelihood that robbers shoot up the casino and steal everyone's money). The statistical tools being used today are just not good enough to evaluate the probability of those sweeping negative events, like the subprime crisis. Consequently everyone will over-estimate the chance of the market going up and up. Taleb bets against all of Wall Street by buying cheap options with extremely small odds. when markets operate normally, he loses small amounts on these options; when the markets lose a lot of money, he calls in the options and makes a fortune.

If you want to learn more about the theory read The Black Swan interesting the whole way through. Read more in this June profile of Taleb in the UK Times, or check out his website.

Liked what you read? I am available for hire.

Stat of the Day

Lee Humphrey Florida's Lee Humphrey, the "other" Florida starter, has played 1180 minutes this season, yet has only shot 13 free throws. The entire season. Florida may lose tomorrow but I think it's unlikely...they tore apart Ohio State in January, they have the experience, they have the efficiency, and they have backup big men.

Liked what you read? I am available for hire.

My thoughts on the Tournament, seeding, and scoring systems.

Penn vs. A&M The story of the tournament so far has been the lack of major upsets. The lowest seed still remaining is #7 UNLV, and many strong teams survived the tournament's first weekend. I watched Penn play Texas A&M live, but missed a lot of the other action. Here are a few of my thoughts on the tournament: 1) If the seeding committee wanted to design the bracket so every higher-seeded team would win every game, they would set it up differently from the way it is now. They would set it up as 1 vs 9 5 vs 13 3 vs 11 7 vs 15 2 vs 10 6 vs 14 4 vs 12 8 vs 16 Which means that the current arrangement isn't as high-seed friendly as it could be, and which means that lower seeds (like Penn) can complain they had to play too difficult of a team, and the committee can say, "The higher seeds won every game, clearly our seeding process was good" and they can both be somewhat right. 2) Has anyone tested to see if the committee suffers from a recency effect in seeding teams (Syracuse gets a 5 last year, A&M gets a 3 this year, Oregon gets a 3) for the tournament? It could be tested with win percentage in the conference tourney vs. performance against the seed expectation. I think win percentage alone is over-rated by the committee. There are strong teams that play tough schedules and lose some games (North Carolina), and teams that get lucky and win more than they should against average teams (Long Beach State). 3) Can the preponderance of higher seeds getting to the Sweet 16 can be explained through the same probability as last year's brackets, or was the chance made more likely by the NBA age rules change? I hypothesize that the disparity between the best teams and the average teams in college basketball was made greater by the presence of players who would have jumped straight to the NBA, but instead played a year of college ball. 4) Tourney pools should give points for correctly picking teams to lose games in the tourney. For example I may pick Wisconsin to lose in the second round to GT, which may be because I think GT is good or it may be because I think Wisconsin will choke against whoever they play. So if Wisconsin ends up losing to UNLV, I think you should still be able to earn points in your pool. Otherwise if you have a team still 'alive' in your bracket vs. a team that you called incorrectly (say I picked Duke to play UCLA in the Sweet 16) even if you think the team that's still 'alive' is going to lose, you should pick them to win cause it's your only chance at getting points. 5) I really enjoy reading Ken Pomeroy's website, if only because they employ statistics and analysis to justify their opinions, rather than talking off the cuff. Here's an explanation of what he does at his website.

Liked what you read? I am available for hire.