Maths Question

[davidevans44]

Hi,

I wonder if anyone could help me with a calculation in Excel.

Say I have data of how each horse in a particular race has previously traded (Pace Data, Previous In Running Lows and Highs etc) and I give each horse a rating of how likely it is to trade say 25% lower than SP.

How would I then factor in the odds of each horse in the race? For example, If the odds on a horse increases/decreases then obviously the odds on other horses will do the opposite to maintain a book of approx 100%. However a horse with odds of 2.0 will have a far greater effect on the odds movement of other horses in the race than a horse with odds of 100 which could for example trade 25% lower than SP with little to no effect on the other selections.

Hope that makes sense. Is there a calculation to weight the probability of a horse trading higher/lower depending on its odds? Or am I thinking along the wrong lines and its not something I need to consider?

Thanks in advance

[Shaung89]

For example, If the odds on a horse increases/decreases then obviously the odds on other horses will do the opposite to maintain a book of approx 100%. However a horse with odds of 2.0 will have a far greater effect on the odds movement of other horses in the race than a horse with odds of 100 which could for example trade 25% lower than SP with little to no effect on the other selections.

This bit certainly makes sense as it is the bread and butter of how market forces work.

I dont however quite understand what you're trying to achieve from the rest - maybe needs a little more detail!?

[davidevans44]

So for example my ratings suggest that 2 horses are likely trade 50% lower in the 3rd quarter of the race, Both selections are joint favourites at a BSP of 2.5. On this instance both cannot trade 50% lower as predicted because the book just would not add up. Now if one had a BSP of 2.5 and the other had a BSP of 20.0 then both could trade 50% lower at the same time.

The shorter the odds the more of an effect that selection will have on the rest of the market when its price increases/decreases. I want to know if there is any way of factoring the odds into my prediction of probability on how a selection will trade.

[davidevans44]

Another way of explaining it is if the 1.5 odds favourite in a race is a held up type of runner and its previous runs all suggest that its likely to drift right out before challenging at the end of the race then it makes all the other runners far more probable to trade lower in those early stages than in a closely contested handicap where the favourite starts at 6.0

So in my opinion I need to add a weighting depending on each horses BSP odds to my ratings to predict whether a horse will trade X% higher or lower

[Atho55]

I`m not sure it`s correct to assume the Book% is around 100% IP as it can deflect wildy from that median. An example..

Book%.jpg

An alternative solution would be to determine how many ticks typically you would expect each Rank or Fave pos`n to move. You would not expect a rank 10 horse to move as many ticks as a rank 1 horse. That might point you towards the offset you can set.

Historical data will give you a pointer to how a runner performs with a measure of ticks between BSP and IPMIn. Set a threshold value and see how it measures against it.

This is for the runners today. A starting point for a threshold value

Threshold.jpg

[davidevans44]

Thats a great suggestion, Thanks for your input

[Derek27]

So for example my ratings suggest that 2 horses are likely trade 50% lower in the 3rd quarter of the race, Both selections are joint favourites at a BSP of 2.5. On this instance both cannot trade 50% lower as predicted because the book just would not add up. Now if one had a BSP of 2.5 and the other had a BSP of 20.0 then both could trade 50% lower at the same time.

The shorter the odds the more of an effect that selection will have on the rest of the market when its price increases/decreases. I want to know if there is any way of factoring the odds into my prediction of probability on how a selection will trade.

There are two possibilities. One, your ratings are wrong (probably most likely ). Two, they will both trade at 1.25 but obviously at different times. I doubt there is a reliable algorithm that can calculate a horses chances of trading low as there are so many factors but a horse's chances of trading low will be interrelated to the others. For example, a front runner may trade low most of the time but in a race with another two horses that force the pace, they could cut each other's throats upfront and set the race up for a hold-up horse.

[davidevans44]

Yes I understand what your saying, The ratings are based on previous races and there are so many variants that they will never be a perfect indicator of how the race will be run, however the odds are an important part of creating such ratings because if there are two similarly priced favourites with totally different running styles then they are very likely to run true to their previous styles and odds move accordingly in opposite directions whereas when comparing horses with higher odds they are not going to effect other runners odds as much.

Its something that I look at manually when trading, I group runners into odds brackets and determine how much they will influence other runners in that same bracket. I tend to concentrate on closely contested handicaps and trade manually so can quickly exit a trade if the race is not run as expected.

Its just that with all this time stuck in the house I have been pondering refining my ratings and taking the odds into consideration with some kind of formula. I know for example from historical data that the lower the odds the more likely a selection is to DOB with around 45% of favourites trading 50% lower in running, this figure is generally reduced as odds increase.

[Atho55]

Would be interested in doing a few comparisons to compare race picks. From 173 picks here yesterday managed 21 winners of varying odds. Usually 4 picks per race. Uk was ok but Ire rubbish.