Myth Buster: Horse Racing Results by Post Position

If you’ve followed horse racing for any amount of time you have run into result summaries by post positions.

A good friend and follower of BTSM forwarded a link to results for the Keeneland October 2012 meet, asking if BTSM takes into consideration Post Position — sending the following link:  (

Results from BTSM Database:

Data Note:  There were actually 76 Races (not 75) on the main track less than a mile.  The PDF from Keeneland looks to have missed the 10th race on 10/27 where ‘LIL MISS SMARTY K’ won from the 2nd post position in a Claiming race on the AllWeather going 7 furlongs.  ‘IRE’ also won from the 2nd post position on 10/26 in the 6th Race (AllWeather going 6 furlongs).


Back to the Analysis

For Main Track < 1 Mile:    The two things that ‘pop’  1) “The Curse of the Turquoise” …the 9th post position going ‘oh’ for 61 and 2) the slight bias in the 1 & 2 post positions.  Therefore — possibly concluding —  maybe the next time you like the #9 horse maybe you should give some pause ……..  BTSM: well (maybe not) … let us take a deeper look.

Breaking it Down:

  • Who comes out of the gate in the 9th post position? (hint: it isn’t always the turquoise saddle cloth)
  • Are these post position results tables something you should care about when handicapping? (Repeatability of Performance)

Before jumping into the analysis (all things being equal) — for a horse to come out of the 9th position there have to be at least 9 horses – which means a relatively large field — which means each horse has a lower probability of winning than a smaller field — just some food for initial thought.

Who comes out of the 9th gate (post position)?

Due to Entries and Scratches, of the 61 horses at Keeneland that came from the 9th gate and only 29 (47.7%) had a turquoise saddle cloth (#9).

1 was the #8, 25 were the #10, 5 were the #11 and 1 was a #12.  So it would be quite an effort to avoid the 9th gate as a manual handicapper.  But if the findings are significant and consistent — working that data into BTSM modeling would prove a useful edge on the market.

Are these post position results tables significant and stable over a longer time frame?

More of an “Expected” vs “Actual” question.  How many expected winners did we “the betting public” expect out of the 9th post position, and how many did it return (chi-square test for the stats geeks following).

‘oh’ for 61 — surprisingly not significant at 95% confidence interval ……. BUT!! — it is close.  Below is a table of how many winners were expected (based on final odds) by position and how many winners were observed – and the difference.

No statistical significance – but what is looking bad is the 1&2 post positions.  The inside horses are not performing.  To determine if there is really something there – we need more observations – robust year or year trends.  Check it out below.

Now we have some data.  637 horses came from the 1&2 post positions from 2009-2012 October, <1 mile on the Main Track.  There should have been around 64 winners and there were only 38.

The ROI for betting on the 1&2 post position for the past 4 years is -49% (compared to a -17% base ROI — track take).  Broken down by year:  -81% in 2012, -32% in 2011, -37% in 2010, -47% in 2009 and for good measure -34% in 2008.

This is statistically significant and holds across purse amount and race type (Stakes, Maidens, etc…).
So – thanks to our friend for forwarding along – you made us Smarter.