Friday, 16 December 2011

One for the future?

I thought I’d post this piece of analysis as I’ve been working on it for a wee while, behind the scenes. Conscious of the fact that the blog is in danger of falling off a cliff in respect of losing readership as I’m not really finding enough new topics to post about. Of course, people come back and follow the bets but I always hoped the blog would be more than just people using it to gain free bets!

Anyway, what I’ve been looking at is trying to devise a TFA ‘super system’. I’m conscious of the fact that people see 20 systems, think it’s too complicated and decide they’d rather follow a simple blog or a simple tipster giving out a few selections every weekend. May not be as profitable as what I’m doing (unlikely to be as profitable I would suspect!) but then again, you won’t need to read endless amounts of posts to understand which bets to follow as they are laid out on a plate for you.

My first idea was to try to look at pinpointing the very best Home and Away bets that appear on my systems but being honest, I could go to town on backfitting results here to suit myself. If you back Home bets that only appear on systems X, Y and Z together, you’ll be thinking you can retire off the future profits when in reality, I’m simply falling into a trap of backfitting.

This season, I have a few systems on the go involving the new systems from algorithm three and of course, I have the pretty useless (so far!) SGM bets on TOX, STOY and STOZ.

Therefore, at the moment, I have 20 systems on the go. What are the results depending on how many times each bet appears?

As you can probably imagine, this isn’t exactly a project to undertake a few weeks before Christmas and with massive databases of results for each algorithm and set of systems. I could easily get lost in data for weeks, if not months here as trying to manipulate over 25,000 bets on all the systems isn’t as easy as you may think as my data for each algorithm isn’t identical.

I won’t bore you all too much with the fact it’s taken a good few hours to get to the results I wanted to find. Of course, going one step further and pulling out data by season, month and so on isn’t impossible but it would involve creating a brand new data set of all 25,000 games and being honest, I don’t have the time. Not even sure it’s a summer job but I reckon I could do it next summer.

Anyway, here’s the results below along with a graph which I’ll explain below.

Simply, the table shows the returns by backing teams depending on how many times a team appears on the 20 systems. For example, only backing teams that appear on all 20 systems will have only generated 68 bets in 5 and a bit seasons. However, you would have made a profit of 33.4pts for H/A and 20.2pts for DNB.

As you work down the table, you can see that the fewer times a team appears on the systems, the less of a return is achieved.

If you graph the cumulative ROI as you work down the number of times appear on a system, then you can see it is a fairly steady trend. As a team drops from 10 systems, to 9 systems to 8 systems and so on, the drop in return is almost linear!

In addition, you can see that as the number of times a team appears drops, the gap between H/A returns and DNB returns narrows a little.

All very fascinating stuff.

What does it mean? Well, one thing I haven’t managed to achieve yet is to build a single TFA system that people with little time can follow. They may only want 3 or 4 bets a week with a simple staking plan. These results effectively show that I can build this single system for next season. I already own some of the best football systems that are available but I think I can add to it by having a single system that people can follow that is easy to follow and simple to track and will save them from having to look through 20 systems to come up with the bets for themselves.

Although it may appear like I’m not doing much at times as there isn’t much to read on the blog, you can rest assured that I’m usually looking at some other way to maximise the edge I have for myself and others following. ;)

The weekend bets will be posted on the blog tonight.


  1. Your selections are almost identical to those of Football Investor especially this week.

  2. Hi Anonymous.

    Thanks for the comment. You won’t be surprised to hear that I already know this if I’m honest. There is massive correlation between TFA and FI bets most weeks. However, from my limited understanding of the FI service, I think that his ratings are based on shot supremacy which as you’ll know if you read the blog, is the very same idea as my ratings. Hence, when two people are both using ratings based on the same underlying factor, I’m not overly surprised that there is a lot of correlation in the bets.

    However, I don’t see it as a big issue. I issue my bets before FI release their bets and therefore, anyone following me has the best chance of achieving best prices. In addition, if you look at the results over the past season and a half, although we may have many of the same bets, our results aren’t exactly correlated. Enough said.