Monday, 18 June 2012

Preview of System E1

It’s interesting. I was looking back to the review of system 6 for some ideas about how to start this post and I quickly realised something about these new Euro systems. I have absolutely no emotional attachment to any of the systems at the moment. When I first developed the UK systems, there was a great excitement around the fact I had developed something that I thought would work. I couldn’t wait to see the systems go live and therefore, over the past 2 years, I’ve grown really fond of each of the systems as the potential has turned into profits.

At the moment, I see these Euro systems as letters and numbers rather than football systems. I’m sure that will change over time and as soon as they go live, they will take on much more meaning but at the moment, I feel a little distant from the systems if I’m honest.  However, I think part of it is maybe the expectations. I’m sure I’m not alone in the fact that I expect these systems to make a profit next season. I think that dulls the excitement somewhat. I feel more pressure than excitement now if I'm honest.

Anyway, here’s a look at system E1. The intention of these reviews is to populate the table which shows number of bets, target ROI, ROC etc. In addition, I’ll try to pick up any trends I see but I don’t want to overanalyse backtested results too much!

What I will try to do is use the elder UK sibling of each Euro system to try to anticipate the drop off in performance we may see once the system goes live.  Hence, in the case of system E1, we’ll use system 6 to try to understand what may happen when E1 moves from backtesting to a live environment. Never going to give a perfect answer but it should help manage expectations and set realistic targets for the upcoming season on each system.

System E1

A reducing number of bets each season but the returns season on season are quite volatile if I’m honest. 2008/9 and 2009/10 are both fully backtested seasons and yet, one ended with an ROI of 9.8% and one ended with 17.2%.

A 12.9% return last season looks low but then again, I’m comparing it to the previous season which isn’t a good idea as I’ve said a few times on recent blog posts. The 3 ‘clean’ seasons produce an ROI of 13.3% overall which is probably the true underlying performance during backtesting I think.

Germany is easily the best league with France easily the worst league. You can see the draw impact on France though where the returns don’t drop by too much as you move from outright, to AH0, all the way to AH0.5.

There has only been one league that has ever lost over a season and that was France last season. If you go back another season, you will only see a return of 2.3% that season, even though the season is probably overstated. Definitely going to be wary of the French league next season on this system (and on most systems I imagine).

41 winning months with 9 losing months historically. Average winning month of 18pts, average losing month of 4.4pts. During the 2008/9 season, the system started with 4 losing months out of 6 which is the worst run historically. Had 2 losing months last season.

Aways had their worst ever season last season overall, with only a 5.3% ROI. Homes had their 2nd best season with an ROI of 19%.

Aways in France are clearly the issue in that league with draws. An overall ROI of 13.8% if using AH0.5 against an ROI of 10.4% if backing outright. Definite trend there and on a season by season view, 4 of the 6 seasons have been better if using AH0.5 on Aways in France.

Aways in Germany look unbelievable with an ROI of 46.9% across 320 bets to date. Wouldn’t be surprised if a correction was due on these!

Homes in Italy look very strong which is what I would have expected I think as traditionally, it is seen as a strong league for Homes.

The overall P&L graph is interesting. Starts well, slows down and then accelerates again. There was a period of 500 bets where no profit was made. That is equivalent to a season. Seen the same things before on the UK systems, just shows that patience will be required at times.

The drawdown graph looks fairly common with no big peaks but plenty of drawdowns around the 20pt-30pt mark. 20pt drawdowns seem to be hit fairly regularly.

OK, in conclusion then, we’re probably looking at 500 bets next season assuming the downward trend continues. The historical ROI of 17% is probably more like 13.3% if we exclude the 3 seasons which may be pulling it up slightly. If we do the same on system 6, the overall historical ROI (2006/7 to 2009/10) went from 17.5% to 15.2% if we exclude the seasons affected by backfitting slightly.  The live results on system 6 has seen an ROI of 5.2%.  Therefore, it’s not impossible that the 13.3% historical ROI may only end up being around 1/3 of this figure when live.  Hence, an ROI of 4.5% is going to be the target on this system for next season.

The max drawdown of 30.5pts would mean we’re looking at a betting bank of around 75pts I think. I’d rather play on the side of caution here given we’re basing a betting bank on drawdowns during backtesting.

So, a target ROI of 4.5% across 500 bets would give a pts profit target of 22.5pts. Based on a 75pt bank, that’s a target ROC of 30%. I would hope the combination of a higher ROI and possibly a higher number of bets than I’ve forecast may mean that this is a conservative view of ROC but better to err on the side of caution here.

No comments:

Post a Comment