Originally Posted by thorns
Its a shame that tinypic has ceased to exist, as some of the data in those images was very handy for anyone new to all this. Was enough that a hack like me could find some dodgy short priced favs to lay in early 2yo races.
I'm avoiding stuff I should actually be doing so I recreated some of the deleted images (i.e. using the same data and not with any of the newly announced chefs de race). People should now be able to easily get the figures necessary for horses from pedigreequery.com, which included the Australian chefs last time I looked. Please be careful if choosing to add new horses to their dataset.
Here is how the CD varies over distance range (NB: races >=1150m and <1250m are collated into the 1200m datapoint, for example). You can see that there's not a huge difference between that of the average winner and the average of the field; maybe a little at each of the distance extremes though that is also where I have less data.
Here is the 95% Confidence Interval of the Winners' CD over distance. In theory, over time, horses contesting a distance with a CD outside of this range should be priced at $20 or greater. It's not likely to be the case in any one race though as it doesn't take into account the abilities of the rest of the field, how experienced the horse is, where it is in its preparation, etc.
This presumes that the CDs of all contesting horses are normally distributed. So a new bonus graph for this post! Here is the spread of CDs for winners of 1200m races (the distance with the most races in my sample). It's close enough to normal but has a slightly bigger right tail, which probably has something to do with the sprint distance being contested + the larger number of speed influencing chefs. Unfortunately the graph for non-winners is almost exactly the same (the peak only very
slightly to the left, as is indicated in the first graph).
Last of all, here's the average tally of individual attributes for winners, varying over distance: