Normal 0

Normal 0

BELMONT STAKES:  Does the Post Matter?

Copyright Robert Crowne Assoc, June 10, 2011

The post position draw for this Saturday's Belmont Stakes took place Wednesday, and final jockey assignments were announced.  Here are the results:

Animal Kingdom was installed as the 2-1 morning line favorite and Schackleford's odds  were set at 9-2, despite Shackelford having beaten Animal Kingdom just three weeks ago in the Preakness.  The morning line odds are set by the track racing secretary, and are the closest modern equivalent to the odds that were set by on-track bookmakers before the advent of parimutuel wagering.  Once parimutuel betting starts to have an effect, Shackleford's draw of the #12 post may work to increase the gap in odds versus Animal Kingdom. 

The inside posts in the Belmont Stakes are thought to have an advantage, and the outside posts are thought to have a disadvantage.  The reasoning is based on Belmont's main track being exactly 1 ½ miles around, and the Belmont Stakes being a 1 ½ mile race.  As a result, the race starts at the finish line, which is only a short distance from the Clubhouse turn.  The proximity of the start to the turn is believed to allow the horses leaving from the inside 1-3 posts to save ground, while causing the horses leaving from the far outside 10-12 posts to travel a longer distance into the turn and exert themselves more than the inside horses in order to gain good position. 

The theory of post-position bias is often supported by a statistic showing that Post 1 has produced 22 Belmont stakes winners, while only 6 horses have won from Post 12.   The number of wins attributable to Post 1 is almost double the number of wins from any other post.  Posts 3 and 5 stand tied for second with 13 wins each.  Post 2 has 11 wins.  The number of wins for horses breaking from Post 12 is half the number of wins from Post 2,3,or 5. 

Based on the above theory, breaking from Post 1 would benefit Master of the Hounds. Despite that, the NYRA Racing Secretary, P.J. Campo, set the odds on Master of the Hounds at 10-1.  Before the post positions were assigned, Mike Watchmaker estimated the odds on Master of the Hounds would be 8-1.  Obviously, Master of the Hounds does not seem to have been given any credit in the morning line for his rail position. 

Perhaps the NYRA racing secretary knows that the post position bias attributed to Post 1 is not what it is generally believed to be.  In fact, the effect of any post position bias has been exaggerated, and the statistics used to support the theory are faulty.

The theory of a post position bias fails to take into account an intentional track bias designed to negate the effect of the greater distance that the outside horses must run.  The track at Belmont slopes inward toward the rail.  Being able to run downhill to the rail compensates for the extra distance that must be run to get there.  The wide profile of the clubhouse turn also lessens the negative effect of the turn.

The problem with the cited statistics is that they are based on every Belmont Stakes race since 1867, including 40 races run at different tracks during the years 1867 to 1904, and 1963 to 1967.  The statistics also include 15 races from 1905 to 1919 that were run in the opposite direction.  Further complicating the situation, in 1958 the track shortened the far turn by 96 feet, and in 1990, to accommodate the Breeders' Cup, the track widened the clubhouse turn.  Finally, the very low number of wins from Post 12 can be attributed more to the fact that many races did not have 12 entrants.  Post 12 can't win a race in which there is no Post 12.

Eliminating the years in which the race was run on different tracks, and the years in which the race was run clockwise instead of counterclockwise, we find that Post 1 has won 16 times in 87 races (18.3%).  Even though that was still the highest number of wins among the post positions, the general distribution of wins in these 87 meaningful races begins to look more like it is attributable to luck than to a bias.  The next highest number of wins in these races comes not from the inside posts, but from Posts 5 and 7, in the middle of the field.  Those posts had 13 wins each.  Post 3 had 11 wins, but, strangely, post 2 failed to follow the pattern of the posts on either side of it.  Surrounded by post 1 with 16 wins and post 3 with 11 wins, post 2 only had 7 wins.  Strangely again, even though posts 5 and 7 each had the second highest number of wins at 13, post 6 between them had a mere 4 wins.  The failure of the win statistics for each post to follow a predictable inside vs. outside pattern suggests that luck, and not track bias is at work. 

The only clear statistical trend that appears in the 87 meaningful races is that odd number posts win much more often than even numbered posts.  Posts 3,5.and 7 produced 37 winners, while posts 2,4, and 6 were responsible for a mere 22 winners.  If we compare instead posts 1,3,and 5 as the odd numbered posts, we get 40 winners from odd numbered posts, compared to 22 winners from even numbered posts.  The difference between odd and even results defies any logical explanation.  The various post position win results over the meaningful 87 years appear to be random, perhaps with a very slight bias to post 1. 

Since 1991, after Belmont widened the clubhouse turn, the inside posts appear to have lost any remaining hint of an advantage.  Posts 1-3 accounted for a mere 5 wins combined in the 20 years through 2010, whichi is exactly equal to the number of wins produced by posts 5-7 combined.  The posts with the most wins from 1991 to the present have been Post 4 and 7 with 3 wins each.  Between those two, posts 5 and 6 produced just one win apiece.  Posts outside of Post 7 accounted for 35% of all wins.  The number of wins for all posts over the 1991-2010 period is very close to random expectation.  The individual win statistics for posts 8-14 have not been considered because those posts did not exist in every race, but the combined statistics can be deduced from the total wins of posts 1-7, and those combined statistics do not present any statistically significant aberration. 

One more factor should be discussed.  There is a predicted 50% chance of thunderstorms for Saturday.  The rail at Belmont becomes very slow and deep when the track is wet.  Jockeys familiar with the track attempt to race all the way out in the four or five path when the track is sloppy.  No matter how much extra distance must be run around the turns, the disadvantage is not as great as the disadvantage experienced by attempting to run in the deep bog at the rail.  Should the track come up wet, the rail horse will experience a significant hardship on the break, and will need to run uphill to get off the rail.