Pegasus World Cup: $12 Million Race, Questionable Timing?


To read more about how TimeformUS Speed Figures are created, click here




Update: Feb 1, 11:45 am

  • After additional review, TimeformUS is sticking with the 2.44 second difference between the Pegasus and Poseidon, giving the Pegasus a final time of 1:46.53. I’m confident that is correct to within a few hundredths of a second.
  • This time would establish a new course record for Gulfstream at the distance (currently it’s 1:46.86 for Lea in the 2014 Donn).
  • I’m going with the same track variant for both races; this decision was unchanged by the re-timing of the race. The pace was slow for the Poseidon, which may contribute to the assessment that the track sped up through the day.
  • Because of the the new timing, the TimeformUS Speed Figure for the Pegasus has been adjusted from a 134 to a 139, matching the best figure we’ve assigned since at least 2013 (Arrogate and California Chrome both ran 139s in last year’s BC Classic).
  • TimeformUS PPs will display a “p” for the slow-paced Poseidon and a “t” to reflect the timing issues with the Pegasus. To learn more about TimeformUS Question Mark Figures, click here.
  • Below: The pace figures listed are for the leader at each point of call.
  • In the re-timed Pegasus fractions, Arrogate gets his final furlong in 12.44 seconds rather than 13.71 seconds.



Update: Jan 31,  8:45am

This supports the notion that the official timing of the Poseidon was correct.  The track appears to be slightly longer than a mile and an eighth, thus the timing should begin after the horses pass the wire for the first time.  The Pegasus appears to be the race in error.


Original Piece: Jan 30th, 10:30 pm

The TimeformUS Speed Figures for the Pegasus World Cup under review, as I have concerns about the accuracy of the official final time for the race. Here’s a preliminary report on the review process:

Background on the original 134 TimeformUS Speed Figure for Pegasus winner Arrogate

There was another race run earlier on the card, the $400,000 Poseidon Stakes for older males, at the same distance as the Pegasus.  The track remained fast throughout the card.  The difference in the official times of the two races seemed reasonable given the quality of horses in the respective races:

  • The Poseidon Stakes was officially timed in 1:48.97.
  • The Pegasus was officially timed in 1:47.61.
  • Trakus is the official timer for Gulfstream and those times are reflected on both the Trakus website and in the charts provided by Equibase.

Using those times, I calculated that Arrogate had run a TimeformUS Speed Figure of 134. This was slower than he had run in his previous two wins but still a standout performance.  A grade 1 race for older males will average about 128 on the TimeformUS scale.  I published the figures for the day and was done with that card.

Subsequent questions arise

On Monday morning, I received a note from noted clocker Bruno DeJulio asking me if I had timed the Pegasus. He mentioned he had clocked the race in 1:46.87 and 1:46.91 on his two attempts.  I remembered seeing some odd things on the Trakus display while watching a replay of the Pegasus so I figured I would download the race and time it using video editing software.  I use this often when something looks amiss with race timing and have written a few articles in the past about races that had problems.  I also downloaded a replay of the Poseidon to use as a frame of reference.

Roll the tape: Upon Review, Troubling Details

Both stakes races were run at the distance of one mile and one eighth, and since Gulfstream Park is a mile and an eighth track, official timing should start and end at the finish line (more detail on this shortly).  This makes timing from video replays relatively simple.  Equibase listed the run-up for the races as slightly different.  The run-up for the Poseidon is listed at 70 feet; the Pegasus is listed at 60 feet.  The replays tell a different story.  The Pegasus clearly starts at a distance further from the finish line than the Poseidon.  It isn’t a huge difference, maybe five feet, but it certainly isn’t 10 feet shorter.  I timed the races using a few different points of reference.  First, I marked the time on the video replay when the gate first began to open for each race.  I then marked when the leader hit the finish line the first time, which in this case should also be the official start of timing for the race.  The race run-up, mentioned earlier, is not timed officially.  Finally, I noted the time when the winner hit the finish line.  All of these are subject to some error, but it is small, perhaps a few hundredths of a second at each.  The measurements and times I calculated from these are listed below:


Initial Verdict

  • The time I calculated for the Pegasus was very much in line with Mr. DeJulio’s clocking, matching his first attempt and only 0.04 seconds off his second attempt. The official times show that the races were 1.36 seconds apart, while my video timing shows the difference was 2.44 seconds.
  • This difference of over a second (1.08) is a huge amount of time for a horse race and very meaningful to speed figure makers like myself.
  • If the Pegasus was 2.44 seconds faster than the Poseidon instead of 1.36, either the Pegasus was faster than originally suspected, the Poseidon slower, or some combination of the two.

Subsequent Testing

I did one final test that I like to do when reviewing race times.  I back the video up from the finish line, by the amount of the final time, to see where the horses were at that point.  They should be at the official start of the race.  Here is where the horses were when backing the video up from the finish line by the amount of the official time:

Poseidon (The horses are beyond their first run past finish line):


Pegasus (Just out of the gate, the horses have yet to reach the finish line for the 1st time):


For comparison, below are the on-screen images using the revised times I calculated–1:49.31 for the Poseiden and 1:46.87 for the Pegasus.

Poseidon (On the finish line for the 1st time):


Pegasus (On the finish line for the 1st time):


At a minimum, one of these races isn’t timed accurately, and it could be both.  It is impossible to definitively tell with the information given.

A Study In Contrasts

I decided to check another race to see if it matched either of the above.  The Sunshine Millions Classic was run at the same distance on Saturday, January 21st, one week earlier.  I did the same exercise, moving the video back from the finish line by the official time to see where the horses were at that point.  The horses are in the identical spot where they are shown in the Poseidon picture listed first, ahead of the finish line.

The oddball is the Pegasus.  For some reason, it is not timed the same as the other two mile and an eighth races I checked.  If it were timed like the other two races, the final time would be more than a second faster than the official time currently listed for the race.

Current Status

The TimeformUS Speed Figures for the Pegasus and Poseidon will be revisited in the near future and will likely be changed. We’ll continue to review this race, and will look to provide an update later this week.


This entry was posted in Data Studies, Race Previews. Bookmark the permalink.

20 Responses to Pegasus World Cup: $12 Million Race, Questionable Timing?

  1. world be free says:

    Buy a cheap stopwatch and get a world champion. Better yet, how about using GSP as coming home in 12.2 geared down sounds very suspect.


  2. Pingback: Trip Notes – January 28 | A Form & A Fedora

  3. John K. says:

    While you’re on the subject of splits and official time – how about turf races that have different rail settings, as those races, whose splits are on the turn, are actually running further than a quarter mile, depending how far the rails are out.


    • The timing is adjusted to account for the extra distance when rails are moved.


      • John K. says:

        If the rail is out 135 feet and the quarter poles are permanent, then the horses are actually running more than a quarter mile for the splits on the turns.

        You need to explain how the timing is adjusted, please.


    • With Trakus, the poles are meaningless. With other tracks using conventional timing, the equipment is set up in different locations for each rail setting. With the rail out, for example, when the horses hit the half mile pole there is more than a half mile to go.

      To test this, watch a turf race with the rails far outside. The times don’t register when horses hit the poles, they will show on the screen before that happens. I hope this makes sense, can get pretty complicated.


  4. Ernie says:

    Outstanding follow-through by you and, of course, Bruno.


  5. Alan Porter says:

    There was definitely something awry with Trakus, as it took two days for the splits to be published, and the “chiclets” were jumping around all over the place (Shaman Ghost jumps backwards and forwards in the stretch if you look). It would be interesting to know whether Arrogate really slowed from 24.1 (12.05 for the furlong pace) for the fourth quarter, to 13.71 for the final furlong (27.41 quarter pace). Keen Ice, who sat off the pace early, and is a famed strong closer is only down as coming home in 13.31 (26.61 for the quarter), and Breaking Lucky and Prayer For Relief, who were near the back as 13.47 and 13.19.


  6. Well I am glad there are people like you who have the time to analyze these posted times and figures they produce. I have always been amazed that starting gates are not routinely placed an exact amount from the first timing post from track to track. Run up times should be exactly the same distance for each track so that the clockings should be started when these racehorses have reached their maximum speed. The mile distance at GP has always seemed to be a problem with timing because there is virtually no run up to first timing post. First quarters in those races are almost always a second slower than at any other track I follow. It almost always has a 2nd quarter that is faster than the first which is not the way US horses are trained to run. Can you and all that do such figures implore and beseech everyone involved to get this fixed???


  7. Nathan Imhoff says:

    Also, I just watched this race side-by-side vs Quality Road’s FL Derby and Lea’s Donn. Going from finish line to finish line, Arrogate beats Quality Road by about 7 lengths and beats Lea by 2 lengths (this is a rough estimate, but it’s consistent). Could this be another Secretariat Preakness?


  8. Nathan Imhoff says:

    Did you have a chance to evaluate split times? I know Arrogate was being eased near the wire, but his final 1/8 seemed very, very slow to me, based on his prior closing fractions. I’m wondering if it was clocked over significantly more than 1/8th? If the timed distance of the race was incorrect, the discrepancy must have come from somewhere, and that’s the logical place to start lookinh, IMHO.



Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s