Average or Median
When doing calculations, whether it is for punting on the horse racing or for other means, do you use the Average or the Median? There is also the Mode which I use for record keeping but not for analytical purposes, maybe I should.
But when it comes to calculating any sort of IMPORTANT stats in my system I use the Median.
The Median is not affected as much by the outliers that occur in horse racing, an outlier like that 100/1 shot winning a race. That champion who when starting as an unknown 2yo wins by 15 lengths. Admittedly, there are not many of the outliers but when you are looking at accuracy, there are enough.
The outliers are more likely to affect your calculations in the early stages of your record keeping. Again, that 100/1 pop is going to affect a list of 10 races more than it will affect a list of 500 races. That is one reason when you commence keeping records that you go back in time and start your records with historical data.
The one area that I use the median for actual punting purposes is taking quadrellas or at least checking the runners I have decided to take or as importantly those I was going to leave out in quadrellas. I love a quadrella and I will readily admit that the one area where I lack proper punters discipline is taking quadrellas. And guess what, I am good at them.
I use quartiles which I found out about when reading about the box and moustache graph. I will go no further on that but my quaddy method is to find the median SP from previous like races and use the quartile each side.
Here is an example based on true figures,
For 1400m maidens, on either a Good 3 or Good 4 track, with prize money now equivalent to $32,000, there have been 183 races I have records for.
The MEDIAN price of the winners of these races is $4.00. BTW, the average price is $7.13.
I go 46 prices below $4.00 and 46 prices above $4.00 and in that leg of the quaddy I take the horses priced between $2.50 and $7.00.
I have 50% of the winners, based on prices, from previous races of this type
I will say though that if there is a short one and you like it, don’t use this method.
And I do not actually count the cells, there are formulas for that. Below the historical list of SP’s for each race type, I have the following 5 formulae,
=MAX(array)
=QUARTILE(array,1)
=MEDIAN(array)
=QUARTILE(array,3)
=MIN(array)
They give you a quick snapshot of the previous details of this particular type of race. They answer, are you backing or keen on backing one of thos rare outliers.
Yes, I am aware that these quaddies will be based on information a while before the second, third and fourth legs will be run. So will any quaddy you take and if you miss a selection due to a late firm or drift, as Ned Kelly said “such is life”.