Fat Tony and the 2010 Baseball No Hitters?
Image from www.simpsontrivia.com
Tonight I read that Matt Garza pitched the 5th no-hitter of the 2010 season. The way the 2010 season started it seemed as if we were on pace to see the total number of no-hitters games fall outside the three-sigma rule of statistics. My initial suspicion was that something had changed about the baseball that was giving pitchers an edge.
I mention the character of Fat Tony in my blog title because it refers to a fictional character Nassim Taleb refers to in his book, “The Black Swan”. Fat Tony is basically your typical street smart, mafia guy who does not know a lot of math, but can easily spot a rigged game unfolding. So I wanted to see if we could say that the occurrences of no-hitters being thrown in 2010 was simply a fluke or perhaps is there something about the baseball, scouting technology, etc. that is giving pitchers an edge? Here are some numbers to consider:
Total no-hitters (recognized by MLB) to date from 1875-2010: 268
Total games played during this period: 196688
Chance of no-hitter being thrown per game = 268 / 196,688 = .0014 or .14%
So using this .14% average, I wanted to see how many no-hitters would be expected in the 2010 season
Total teams in MLB 2010: 30
Total games played per team 2010: 162
Total statistical games for 2010: 30 x 162 = 4860
Total expected games played for 2010 = 4860 /2 = 2430
Expected no-hitters for 2010 = .0014 * 2430 = 3.31103067
So already we are 2 games over the expected number based on the historical average. But what about calculating the standard deviation? So here again, I got a little lazy as I did not want to calculate the total games played each year, the no-hitters thrown that given year and then find the standard deviation across this data set. Instead I just took the last 19 years of data and included the projected no-hitters thrown this season based on the current trend to run some calculations.
| Season | No Hitters | |
| 1990 | 7 | |
| 1991 | 6 | |
| 1992 | 1 | |
| 1993 | 3 | |
| 1994 | 3 | |
| 1995 | 1 | |
| 1996 | 3 | |
| 1997 | 1 | |
| 1998 | 1 | |
| 1999 | 2 | |
| 2000 | 0 | |
| 2001 | 2 | |
| 2002 | 1 | |
| 2003 | 2 | |
| 2004 | 1 | |
| 2005 | 0 | |
| 2006 | 1 | |
| 2007 | 3 | |
| 2008 | 2 | |
| 2009 | 2 | |
| 2010 | 8 | (projected) |
The projected value of 8 for 2010 is arrived by calculating the total games currently played in 2010 as of the time of this article (1477) and then using the ratio to a full season (2430). Or 5 / (1477 / 2430) = 8.23.
Now already there were other problems that cropped up. One thing was the strike shortened 1994 season where MLB effectively canceled all games on August 12, 1994. Also, should some revision have been made to the number of no-hitters during the peak of the steriod-era (note that there were only 9 no-hitters thrown from 1999-2006)? Finally, what to do with the postseason games? The total number of games played there also need to be factored in. In the end, I decided not to make an adjustments for these three items. So using the data I found the average during the past 20 year period was 2.38 no-hitters per season. The standard deviation is 2.16.
Alas, it would appear that Fat Tony has not rigged the baseballs. Even though we are currently on pace to see an all-time high (excluding data from the 1800s) for no-hitters thrown in a season, we are still within 3 standard deviations as 2.38 + (3 * 2.16) > 8.
