The latest e-mail I received from RA says of the foundational strategies,
"I had to make sure that: They showed an advantage straight out of the box."
I'm not sure what is meant by 'advantage', whether it means over the market, over random selection, over competing products. Can this be clarified and has it been achieved?
Both questions were answered in the second e-mail which arrived this morning. 'Out of the box, the foundation strategies using 'two or three' ratings have an A/E greater than one. It would be good to know the SR, A/E and Chi2 figures for these strategies as this is the baseline the strategies would regress to but which you are seeking to better.
Perhaps @michael-clarke might choose to follow one of these for illustration.
We're aiming to give strong starting points on which you can build. The SR and A/E are available, but not the Chi2, and where it can be we show them for both win and place markets. When the system builder is released the Foundational Strategies will be able to be added into the system builder (where Chi2 will be available) and can then be adapted further.
I have been looking at foundation strategy 1 and I can confirm that for 2019 the AE was an impressive 1.07.
However when I analysed the data for years 2014 to 2018 the AE was only 0.98, bringing the overall AE to less than 1.
Personally I like to have at least 4 years data when producing this sort of strategy and 1 year of test data.
There may be some hope in that 2yos have an AE of 1.07 over the 6 years and class 1 races have an AE of 1.06.
I have analysed the factors used in the foundation strategy 1 for the years 2014 to 2019 and the strategy that I have ended up with is :-
Jockey SR > 15% and < 20%
HRS Win % < 7%
Forecast Odds > 15/8 and < 28/1
Trainer SR > 4%
I have used forecast odds so that selections can be made before the race rather than having to wait to near the off to see what the SP is.
The results over the 6 years are :
1,423 winners from 7,796 runners (18% strike rate).
ROI to proportional stakes 9%.
The AE has been at least 1.07 in each year.
Limiting the selections to SP favourites produces an AE of 1.14 with a SR of 37%.
This does look like quite a good strategy although I haven't tested it on any test data.
I will do a similar exercise on 2014-2018 and use 2019 as test data.