Today has been extremely productive. I have unintentionally wasted this week trying to build various profilers with no real direction or structure to them whatsoever. I realised I had made mistakes with the code which messed up the calculations (e.g. assigning the weighting to LTO1 twice which screwed up the entire dataset). Put it down to exhaustion, inexperience working with code, over excitement, whatever, etc.
Back on track again this morning and with a renewed focus I went right back to the start and built basic but course and distance specific profilers for all 31 turf tracks and the 6 all weather tracks (I was keen to see how they performed). It has taken me 19 hours or so and the delight when I finally made it to York was immense!
Each profiler was backtested to 2018 and surprisingly the majority of them performed well given the standardized nature of the profilers. Overall Nottingham, Doncaster and Chester were the early standout stars. I haven't done any tweaking to any of them yet as I think I have been falling into the backfitting trap all week without realising it. Plus today was about getting all the courses up and running rather than fine tuning.
These are now the official v1.0 base profilers which I will build on over the coming days and weeks. The code has been checked three times and then some so is 100%. To give you an example of the data I have available to me when backtesting we will look at our performance at Chester as it is one of the standout performers and we have selections there tomorrow.
Looking at Chester as a whole rather than breaking it down into individual distances, since 2018 the stats from the profiler are as follows;
Selections: 105
Winners: 48
S/R: 46%
P/L @ SP Level Stakes: 180.60pts
Average SP: 6.0
Longest Winning Streak: 6
Longest Losing Streak: 6
Place S/R: 67%
I think you'll agree those stats are impressive by anyone's measure. It made profit every single year with 2022 being the standout with over 100pts @ SP (I don't record BSP in my data just ISP). The profiler is most effective at the 5F, 10F and 12F distances.
What is exciting is that these margins could all be easily improved with early and or exchange prices and an aggressive staking plan. Also looking at the profiler and the data there are things that could be tightened up for sure which would have a positive effect on P/L and S/R. Plus knowing our longest losing streak allows us to tailor our stakes to suit... Fine tuning and digging deeper into the data will follow over the coming weeks as we develop each profiler and turn them into profit machines.
Ok so the profilers produced 11 selections from 10 races for today. The profilers for York are the worst performing of them all so exercise caution if you are looking at these races. I have ranked these in order of performance i.e. Chester has some of the best performing profilers followed by Leicester, Bath, etc.
Chester 15.55 - Capone @ 2.5
Chester 17.00 - Haarar @ 5.0
Leicester 17.55 - Tees Comet @ 41.0
Leicester 18.25 - Crime Fiction @ 9.0
Leicester 18.55 - Baileysgutfeeling @ 1.6
Leicester 20.25 - Stintino Sunset @ 5.0
Bath 16.45 - Cape Cornwall Rose @ 4.5
Sandown 14.50 - Lakota Blue @ 34.0
York 15.40 - Pure Angel @ 15.0
York 17.25 - Good Earth @ 7.0
York 17.25 - Rathbone @ 18.0
This is the targeted race specific approach I should have adopted from the start but instead I got distracted with all the new shiny lights and buttons and went off on a tangent that has brought me full circle back to here. Out of interest I ran the profilers for the races yesterday and we would have been around 9pts in profit for the day (4 winners and 1 place from 8 selections).
As a side note I have played the two shortest prices as a double to round the stakes to an even 10. With regards to my campaign the above is how I will be moving forward from here but I will take the weekend to continue to test and work on things but also to recover. I have been pushing far too hard lately and am exhausted. Been at this today since 5am or 21 hours ago.
However with all the base profilers now built the system is fully up and running. This is where the fun really starts. From Monday we will go for it, we will target each race in specific detail, tweak the profilers ahead of each meeting, get as much early price value as we can, win market only, go big or go home.
A discussion will need to be had on how to best stake these selections. The backtesting shows the longest losing streak since 2018 happened at Ripon and lasted 12 races. Remember though that we had winning bets elsewhere during this time. The point I am making is you could stake 8% of your bank and not go broke. Not saying I am considering that but food for thought and hopefully some discussion tomorrow.
Thanks,
Darth Niloc
I stand by my original comment, you're a mad man haha
Any how, good to see you back so quickly Darth, you know how to put the hours of work in (while missing work I assume lol) take some R&R over the weekend.
With still being very much the novice to all this, I can't really comment on the best staking strategy. My bets are still very small, I'm talking £1-£5 bets, the wife would kill me for betting and losing too much. So while dabbling for this first year it is safer to remain with smaller stakes for myself, until I've learned more about the sport.
Great enthusiasm and your projected stats are what dreams are made of, get any where near them come end of season we will all be drinking Pina Colado’s on a Caribbean beach at Christmas.
May all dreams come true.
Good luck
I meant to add 30% is our key target here. All profilers have an expected minimum strike rate of 30% baked into the numbers. It should be achievable as even the poorly performing York still has a strike rate of almost 36%. The longest losing streak there out of interest since 2018 is 7 races (or 14% stakes 🤣).