When Avalon Hill acquired 3Ms collection of sports games, it published for a few years its All Star Replay Magazine, which on several occasions featured additional race program cards for the popular Win, Place, & Show horseracing game. One such race program (in Volume 3 Number 2) was for the 1920 Travers Stakes, which featured one of the greatest horses of all time, Man O’ War. In 1920, Man O’ War was a perfect 11 for 11 on his way to being named, along with Babe Ruth, co-athlete of the year.
In the actual race at Saratoga for that year’s Travers Stakes, Man O’ War only faced two opponents, Upset and John P. Grier. But to make a suitable six-horse program card for game play, the Travers was filled in by three other horses that faced Man O’ War in other races in the same time frame at a similar distance.
Of particular note is the entry in Gate 6, a horse by the name of Hoodwink. This horse is listed at 100-1 odds. Hoodwink did indeed race Man O’ War head to head. He was Man O’ War’s only challenger in the Lawrence Realization, where he carried in reality the 100-1 odds and lost by an estimated 100 lengths while Man O’ War set a track record.
The Win, Place & Show program card, created by racing program maven Patrick M. Premo, shows Man O’ War with a Class value of 78 for the 1 ¼ mile distance, six points higher than his next strongest competitor. Man O’ War also has the favorable number 7 bonus. But look at poor Hoodwink – a Class value of a measly 50, and a bonus number of 12 to boot.
* * * * *
Several years ago I created a WP&S race simulator. This simulator runs individual races based on random dice rolls (or pseudo-random, it's a computer program after all), movement of horses by running strength in turn order, limited movement in turns per the actual track, passing lanes, use of apprentice and veteran jockeys, and so forth. A full simulator of all aspects of a given race.
Feeding in the parameters of a race, I would run the simulator thousands of times and record the results – who finished in what place – then I would sum up the times each horse came in 1st, 2nd, etc. Using the original WP&S race program, for instance, these were my results for Race #1:
[click on image to enlarge]
For each horse, there are first five columns of raw compiled finishes followed by the finishes by percent in all races. For example, horse #4, Lucky Old Sun, the favorite in the first race of the original program, came in 1st (for the Win) 2,156 times out of 5000 races, or 43.1% of the time. 3.8% of the time Lucky Old Sun would come in last.
The last two columns are used for a composite strength score and a value I calculate for use in creating odds.
* * * * *
I also compiled statistics for the 1920 Travers Stakes card featuring Hoodwink.
So how did Hoodwink fare?
In this case, I ran 10,000 iterations. Based on these results, I figure that Hoodwink’s 100-1 odds were very generous. I would say 1000-1 would be more in the ballpark. Man O' War dominates with over 63% of the wins. Hoodwink is horse #6 and comes in dead last 96% of the time.
You may wonder, but how does even the lowly Hoodwink finish ahead of Man O' War a few times? How does Man O' War come in last 12 times of the 10,000 races?
I investigated one of the races where Man O' War came in last. After the first move, Man O' War (moving last that turn) was boxed in behind three other horses. But then it was to move before the horses in front of it and lost all of its second movement. this occurred once more in the turn where it lost movement being stuck behind other horses, so it lost much of its movement over the course of the race. Different movement choices (than what the computer made) could have helped Man O' War in that game, but the results show that even the best need to avoid being boxed in along the rail.
But truly Hoodwink is the worst horse ever to show up in a realistic game simulation.
- Last edited Fri Feb 9, 2018 2:36 am (Total Number of Edits: 9)
- Posted Tue Jan 16, 2018 10:30 pm