# PIV is better than Impact Value

In my last post I talked about how you can use impact value to analyse the strength of pretty much any factor, rating or selections in any sport. Today I want to share with you an even more powerful way of doing this.

It is called Pool Impact Value or A/E. I personally prefer Pool Impact Value, a term made up (I believe) by US handicapper Dave Schwartz. In the UK it is more commonly referred to as A/E which stands as Actual / Expected.

PIV or Pool Impact Value makes much more sense to me and so is the terminology that I choose to use!

Calculating the pool impact value is done in a very similar way to calculating the impact value except this time we also include the pool, or in the UK the final or SP odds.

This creates a far more accurate figure because we know that usually the odds are a very accurate representation of a horses chance of winning.

So how are we going to calculate this figure?

You will be pleased to hear that it is much easier than you may expect. We are looking to calculate…

Winners / Expected Winners

The key is in how we calculate the expected winners and when calculating the PIV we used the odds to guide us.

To calculate the expected winners using the odds we use the simple sum…

1 / decimal odds

Decimal odds are the kind of odds you see on Betfair. They are figures such as 2.50 or 3.75 etc… Odds such as 3/1 and 5/1 are called fractional odds and you don’t want to use these with this method.

So if we have odds of 3.75 then we would do…

1 / 3.75 = 0.27

This actually gives us the probability that the horse is going to win the race based on the odds. A probability is the same as a percentage but divided by 100. So 0.26 is the same as 26%. In other words, the odds are telling us that this horse has a 26% chance of winning. We know that final odds are a very accurate estimate of a horses chance of winning and so we use this to calculate our expected winners.

In the sample of runners that we have taken from a particular tipster or rating we perform the sum…

1 / decimal odds

…for each of them. Then we add all of these decimal figures together and the final number will provide us with an accurate assessment of the expected winners based on the odds.

To create the PIV we simply now…

Winners / Expected Winners

…take the winners in our sample and divide it by the expected winners that we have just calculated. The final figure is your PIV or Pool Impact Value.

This works in the same way as the impact value in that anything higher than 1 means the horses are winning more than average and anything lower than one means that they are winning less than average.

The difference is that this is a much more accurate and powerful technique to use than impact value. Now you can go away and calculate whether your favourite selections are winning more or less than the average!

Regarding the PIV, I would like to see a working example for clarification purposes. Thankyou

No problem Jeff. If we have 5 selections with odds of…

5

12

3.75

2.50

6.2

We then do the sum of 1 / Odds for each of these…

1 / 5 = 0.2

1 / 12 = 0.083

1 / 3.75 = 0.27

1 / 2.50 = 0.4

1 / 6.2 = 0.16

We sum together the results of each of these calculations…

0.2 + 0.083 + 0.27 + 0.4 + 0.16 = 1.113

This means that we would have expected to have had 1.113 expected winners from these 5 runners.

If we actually had 3 winners then we would now do the sum…

3 / 1.13 = 2.65

The PIV for these selections is 2.65. Obviously you would need far more selections to get an accurate figure but I hope that clears the process up.

in have had 5 bets

priced

4=0.25

7=0.14

13=0.077

4=0.25

3.5=0.28

total 0.997

they all won

5/0.997=5.01

whats the 5.01 mean

A PIV of 5.01 means that your selections are 401% more likely to win than the odds would have predicted them to. Of course you’ll need a bigger sample than five to get an accurate figure for the method though 🙂

Hi Michael, if you as looking at, say, the PIV for Previous Course and Distance winners, is it not the case that the odds of winning may be lower precisely because of the fact that these horses have this characteristic? Cheers, Gavin

I’m not sure I’m following Gavin. The PIV will use winners vs expected winners. The amount of expected winners is based on the odds of those horses, the SP odds are highly accurate but based on all the available market information. By comparing the amount of actual winners who were previous distance winners against the expected amount of winners based on the odds we can see how much of an impact the previous distance has made and whether it’s a positive or negative impact in this situation. I hope that clears it up.

hi Michael. So thinking on a macro scale, if you took a particular rating, for arguments sake, let’s say 5278 and let’s say I looked and found that a 5278 score of 4 seemed to win the most; no idea if this is so, but just for the exercise. Let’s look at maybe 5 years of data; so I take the BSP of all runners in the given time and convert each to a decimal to show their expectancy to win. I then add all of these together and come up with a figure labelled Expected Winners. Then simply count the actual number of winners and divide expected winners into that. If the figure is above 1, then that particular rating could be said to achieve a positive result. Could be quite useful in analysing which ratings to couple together. Do I have that correct?

Pretty much, although it doesn’t necessarily mean it will be profitable, it means that those horses win X% more than would be expected by their odds 🙂

Just further to that, I said BSP, but I assume you want to be using just the SP result, as the BSP odds can be vastly inflated at times. Is it better to be using SP over BSP for calculation purposes? Thanks for your reply

Just further to that, I said BSP, but I assume you want to be using just the SP result, as the BSP odds can be vastly inflated at times. Is it better to be using SP over BSP for calculation purposes? Thanks for your reply