Using statistics in context

King Yao is the author of Weighing the Odds in Hold‘em Poker, and Weighing the Odds in Sports Betting. He uses his experience from making millions in financial derivative markets and translates it into gambling. Since he left his trading position in 2000, he has been playing poker and betting on sports. He travels to Las Vegas frequently, especially during football season.

Context of statistics

Statistics should be used in context of the teams, players and game situations; sometimes they can be misleading. Rushing yards for a running quarterback can be one of these misleading statistics. Here is a hypothetical example of a handicapper looking at the Atlanta Falcons.

Yards per carry is a key statistic in the NFL. A favorite angle is to bet on an underdog that outrushes its opponents in yards per carry. As you are looking through the statistics, you find a game that fits this angle. The Falcons are a 5-point road underdog to the Patriots. In yards per carry, the Falcons are ahead of the Patriots by 4.1 yards to 4.0. You decide to bet on the Falcons since they fit your angle.

While it is true that the Falcons out-rushed the Patriots by a little, are there mitigating factors that change the way the statistics look? If you dig deeper into the details, you would find that the statistical advantage that the Falcons hold in rushing the ball may be due to the fact they are so poor at passing the ball. Their hypothetical quarterback -- let’s call him Davis -- is great at running the ball, but he is not an accurate passer. Too often he runs the ball instead of throwing it. While Davis may gain yards on the ground and make the rushing statistics for the Falcons look good, he is doing it at the expense of the passing game. Another quarterback in his place won’t run for 80 yards per game; instead, the other quarterbackmay pass for 125 more yards than Davis. Instead of completing a 15 yard pass when it is third down and 14 yards to go, Davis is scrambling and gaining 8 yards, enhancing the rushing statistics but falling short of the first down. Thus the rushing statistics for the Falcons are deceiving; their rushing numbers should be viewed differently due to their unusual quarterback talent. Going into the details of what the numbers represents allows you to understand that this is not a normal case for this specific angle. Now that you know the details, you can make a better assessment of whether you want to bet this angle in this game.

Weed out the Luck

Randomness and luck play a big part of outcomes of team sports. Such sayings as “baseball is a game of inches” and “footballs bounce in strange ways” reflect how luck can play a big part in the final scores of games. Handicappers should try to weed out the luck out of past results and focus on statistics that reflect true ability rather than luck. Handicappers that are able to do this have a better chance of succeeding. Here is an example in baseball.

Sabermetrics is the science of understanding baseball through statistics. More than any other sport, people love to dive into baseball statistics to glean the truth. Over time, research has shown that some popular statistics have a high degree of luck associated with them. A pitcher’s win-loss record is one of those statistics. A pitcher’s win-loss record reflects not only his performance, but also the performance of the fielders and hitters on his team. If he pitches great, but his team does not score any runs, he can’t win the game. If he gives up 5 runs while his team scores 6, he gets the win, but it was luck.

The media rely too much on a pitcher’s win-loss record when deciding their Cy Young Award votes. Good handicappers have the ability to weed out the luck from past results.

This is part of an occasional series of articles.

Excerpted with permission from the e-book version of Weighing the Odds in Sports Betting by King Yao, edited for this format.