Regression Backing - A Collection of Systems


First of all, what do we mean by regression? It's not a horse that's getting poorer as it leaves the peak of its racing prowess but it's a reference to how we generate our ratings. You ought to be aware that we produce the UK Horse Racing ratings purely on advanced mathematical techniques; namely calculus and regressional analysis.

Therefore, a regressional horse is one which stands out in one, or more, of two areas; that of distance and that of going. So, a horse with positive regression on the going can be expected to like the going and the same with the distance. Conversely, a horse with negative regression figures can be expected to dislike the going.

There seem to be no end of various systems involving regression horses, these are the ones which I am aware of and if there are others then please send them in.

GB Stakes

Back any horse with non-negative regressions where the total of the two regression figures is two or more. If one avoids the Flat and NH Flat then one will do even better.

Going Regression

In handicaps, a Going Regression of 3 or higher is, quite simply, profitable.

With just this simple use of the Going Regression with over 1,400 handicap runs we're given a two hundred point profit (after BF Commission) giving us a strike rate of 11% and an ROI of 13.5%.

This example is just with one simple filter and with the Going Regression figure plucked out of the air and fed through the Data Analysis Tool. Other investigations will, more than likely, find other angles. But this is the first investigation which, quite literally, sprung to mind.

Distance Regression

In Great Britain (i.e. excluding Irish racing), a Distance Regression of 3 or higher is, again, profitable.

From this initial starting point one could look at non-handicaps only, or one could look at all races other than the All Weather or simply look at low Race Class races. All of these would simply increase the profitabily of these.