Plenty To Report

What a week this has been. Three weeks ago I was contacted regarding a superb job with my old company and after a few conversations I have decided to go back. I've been with the current business for just under 5 months and suffice to say I wouldn't be making a move so quickly unless it was well worth my while. It was a tough decision to make in the end and it boiled down to gut feel. I always trust my instincts in these situations. The nature of the industry I'm in means that I have been put on garden leave with immediate effect, which gives me another couple of weeks' holiday before I start in my new role. It also gives me some time to continue my systems research, and I'm squeezing in a trip to Epsom on Saturday!

The research is coming along pretty well. I've actually got 3 systems now in test, all based on completely different ideas. One of these, System 2 (I've nicked Graeme's numbering scheme from TFA!), concentrates on double figure value bets and has thrown up winners at 38.73 and 14.50 from 21 bets so far. I am really very keen on this one. System 1, based on a horses for courses idea, is showing an ROI of around 11% after 64 bets and System 3 shows an early ROI of 23% from 118 bets. None of the systems can be back-tested unfortunately, so I need to record results under live testing for some time to become before any conclusions can be drawn. Hopefully these early trends will continue over the weeks to come and I will then be able to start putting selections on here for those who are interested in following them.

Finally, I have decided to add Cassini's XX Draw/Unders service to my portfolio. Again, this provides excellent value for money and the results speak for themselves. Things are beginning to shape up.

End Of The Holiday

As the sun sets on a fantastic week's holiday I thought I would map out my plans for the next few weeks and explain how I intend to update the blog on the various systems I will be introducing to my new portfolio this year.

Before I do that, System 1 performed pretty well today with a profit of 12.56 points from just 6 bets. I also made an error in yesterday's calculations in my hurry to update the results this morning. The profit was actually 0.95 points for Saturday from 23 bets. So, after 29 bets the profit stands at 13.51 points (calculated at Betfair SP after a 5% commission deduction), an ROI of a healthy 46.58%. I haven't applied any filters at all to these selections yet and there are some obvious ways of improving the performance using causative reasoning, rather than back-fitting. As an example, at Chester yesterday the effect of the draw wasn't taken into account, a key variable that simply must be considered at the track.

Once back at work I probably won't have time to update the blog every day with selections. I will try and figure out a way of doing so, as readers will hopefully gain some benefit that way, but with only trial systems being developed at this stage I wouldn't recommend betting anyway for now. My initial plan will be to record the performance daily and publish a weekly review here on each system. At present I only have System 1 on trial and that is going to take me quite a while to build into something workable. I do have a few other ideas bubbling around and I will attempt to model those in around a month's time. 

Meanwhile I am preparing for the new football season with The Football Analyst service and I've been sent a couple of other systems/methods to consider which will take me some time to assess. I do feel very positive about life at the moment and hopefully the blog will begin to take shape over the next 2/3 months and provide readers with some useful content. In the meantime, bear with me!

Sunday Selections

Not a bad day yesterday with the unfiltered results, here are today's selections for System 1:

14:15:00	Kelso	Keeneland (IRE)
14:15:00	Kelso	Cavite Eta (IRE)
14:25:00	Fakenham	Pheidias (IRE)
14:55:00	Fakenham	Phoenix Des Mottes
15:50:00	Kelso	Badger Foot (IRE)
17:00:00	Kelso	Hawthorne Bay
17:10:00	Fakenham	Rebel Flag (IRE)

System 1 Trial

I've managed to build my first system and it goes on trial today. There are multiple selections in some races as I have not applied any filters at all yet. I still have lots of work to do on this, so please don't back any of these, but I am encouraged by the long-term stats:

14:00:00	Haydock	Last Shadow L 8.4
14:00:00	Haydock	Moidore L 4.03
14:30:00	Haydock	Trail Blaze (IRE) NR
14:30:00	Haydock	Forest Row NR
14:55:00	Chester	All Or Nothin W 4.1
15:00:00	Haydock	Bated Breath W 3.27
15:00:00	Haydock	Masamah  L 8.1
15:00:00	Haydock	Swiss Dream L 28.1
15:30:00	Haydock	Whos Shirl L 27.24
16:05:00	Haydock	Anderiego  L 8.4
16:35:00	Chester	Knowe Head L 4.71
16:35:00	Chester	Fazza L 9.93
16:35:00	Chester	Im Super Too L 11.6
16:40:00	Haydock	Bawaardi L 39.03
17:15:00	Haydock	Hometown Glory L 11.9
17:15:00	Haydock	Three Am Tour L 7.0
17:15:00	Haydock	Brocklebank L 38.0
17:45:00	Chester	Green Park W 5.11
17:50:00	Haydock	Cono Zur W 11.32
17:50:00	Haydock	Thackeray L 6.07
17:50:00	Haydock	Fastnet Storm L 20.0
17:50:00	Haydock	Postscript (IRE) NR
18:10:00	Newbury	Resplendent Light L 10.7
19:00:00	Lingfield	Billy Red L 27.44
20:25:00	Newbury	Eagle Nebula L 10.0
20:55:00	Newbury	Emman Bee L 3.85

Total +0.95 points after 5% commission from 23 bets. (Edited 26th May 19:48)

Today's Racing

Just for a bit of fun I've put up some value bets for today below. I've noted the current prices and will compare to SP tonight. I'll see how they do and test out a couple of theories buzzing around my head:

2.00 Fiducia 5.9 3rd of 6 BFSP 5.6 
2.35 Hey Fiddle Fiddle 9.2 8th of 8 BFSP 7.6 
2.55 Blue Clumber 10.5 7th of 12 BFSP 14.51 
3.20 Warden Bond 7.6 10th of 10 BFSP 8.6
3.30 LAY Desert Philosopher 3.5 6th of 8 BFSP 2.5
3.45 Perennial 7.0 3rd of 4 BFSP 5.99
4.20 Ahtoug 5.1 5th of 8 BFSP 4.67 
4.40 Mlady Eliza 10.0 3rd of 5 BFSP 16.33 
5.40 So Is She 8.6 2nd of 8 BFSP 5.8
6.25 LAY Amuse Me 3.0 1st of 4 BFSP 2.02 
6.45 Edas 13.0 9th of 12 BFSP 16.41
7.05 Hellolini 9.6 7th of 12 BFSP 14.0 
        La Sylphe 17.0 6th of 12 BFSP 5.49 
8.15 You'relikemefrank 10.5 6th of 10 BFSP 12.01


Edit 23:06 Friday: Nothing at all from that lot, despite a massive gamble on La Sylphe in the 7.05 race, but this evening I did stumble across something potentially decent which I tested to low stakes with good results. To be clear this was an outright backing idea (one of those buzzing around this morning) and not system related, just an experiment based on something I've spotted:

Nice little bonus birthday present!

Full Circle

I'm in the middle of my first week's holiday since returning to the corporate world and I've certainly picked a good one weather-wise. This time-out has given me a good opportunity to reflect on my decision to quit full time trading at the end of 2011. I have to say that I have no regrets whatsoever, in fact I am much happier now than I was 12 months ago. I have a rejuvenated social life, I am being recognised for what I'm achieving every day and I don't need to worry about making the odd bad decision as the money still lands in my bank account at the end of every month. More importantly I am really enjoying this week off as I value my free time now much more than I did last year. To those contemplating a job as a full time trader I would urge you to think carefully about all aspects of the decision. It's easy to focus on the idea of the dream lifestyle, but it can be incredibly lonely and all-consuming and those downsides have to be considered.

So, what next for the blog? Well, after another chat with JP, who always seems to inspire me to achieve something new and exciting, I am working on a brand new project. Trading is no longer a realistic option for me - apart from the odd NFL game through the winter - so I have decided to try and build a small but perfectly formed portfolio of systems to follow with outright bets. I have thus gone full circle and will be returning to where I started in my pre-trading days, with one subtle but vital difference - I won't be paying much in the way of subscription costs this time around. The objective will be to build a little retirement fund over the coming years. The key will be to avoid any form of pressure with trying to get bets on and obtain value prices, so most of the systems will be my own. I need to be largely in control of my own destiny with the day job to work around and family time at such a premium. However, the first portfolio entry will be fellow blogger Graeme Dand's excellent The Football Analyst service, which at £150 for a full season is stand out value for money given the track record of Graeme's systems over the last 2 seasons under live testing.

In terms of my own systems I have nothing more than a few ideas and a blank sheet of paper right now. I have just signed up to the Dataform horse racing service for 5 weeks (£45) and I have started a massive data research project this morning. I am using the Dataform Systems Analyser which goes all the way back to 1st January 2007, so having learned from my last experience I will be ensuring that every system has a huge number of data points tested before a single pound is wagered. Once I've pulled together a reasonable number of systems my intention will be to publish selections on a daily basis here. I hope that will help a few readers make a few quid along the way. In fact if anyone has an idea for a horse racing system that they'd like testing I'd be more than happy to assist. Either leave me a comment or drop me an e-mail to mark@horseracingtraders.com.

Fooled By Randomness

Imagine this scenario: You spend a few days producing some data analysis on 1000 horse races and the results are promising. You take a second batch of 1000 races and the results are equally promising, in fact they are actually a shade better. You then ask a trusted friend to pick a random period for a third batch of 1000 races and true to form the results continue to prove very profitable. Would you be willing to invest your money in the method that produced these results? Well, if you were, then you would be disappointed. Probably not as disappointed as my brother and I were when we discovered the performance of the method we've been working on when analysed over 29,000 races.

Further to the last blog post we managed to find a fellow bettor/trader/bot guru who agreed to help us automate the strategy. We took a leap of faith and without an NDA or any such legal protection we shared our thinking with Mark and that was a great decision. He was absolutely brilliant throughout the process. A couple of weeks ago he managed to find a way of testing our method over every horse race since 2008 and despite those three early successful periods the overall results were devastatingly poor (a loss of 4600 points over 29000 races). We rapidly came to the conclusion that we had managed, by pure chance, to have selected three separate test periods that were all showing excellent profits in the midst of a long losing period. Balls.

We have learned, without investing a single penny, that when analysing data you can never really have enough. 3000 races were nowhere near enough in this case. The good news is that thanks to Mark we didn't lose anything, nor did we attempt to back-fit the results in any way to show a theoretical profit. We also found a trusted guy who was true to his word and worked incredibly hard to produce a beast of a spreadsheet. The drawing board is firmly back in the room, dusted down and awaiting the next idea. The only problem is that I think my brother and I have run out of steam. We have decided to take an extended time-out from data research. This time next year we probably won't be millionaires.