Thursday 3 January 2013

Half Season Review & Live Results Update

As promised on the last post, I have taken the time to write a review of the season to date. In addition, I think it’s also important that we take a quick look at the live system results so we can put this season into some sort of context. It would be easy to only look at half a season and think some systems may be the best systems ever and other systems may look absolute rubbish but we need to look at as many live results as we have to give some form of context.

As most of you know, this project started over 2 seasons ago now, so this half season is only 20% of the overall live results, so basing future investment decisions on 20% of the data in some cases when we have the other 80% of data available to us would be folly. I’ll touch on this point once I finish the half season review.

So, how have the results been so far this season……Here they are here in a nice little table I’ve used before on the blog.


There is obviously a lot of data in this table as I’ve said before but I think it’s best to look at the high level figures before drilling down into each set of systems.

At a high level, there have been 5,818 system bets this season to the end of December. I expect to proof around 14,000-15,000 system bets this season and in particular, the Euro systems are not even 1/3 of the way through the season yet, so these systems will have many more bets in the second half of the season than the first half.

If simply backing 1pt on every system bet that has been generated, you would have achieved a profit of 466.5pts. This is an ROI of 8%.

This is slightly below where I aim for at an overall level (10% would be this figure) and long-term, the system bets before this season had achieved about 8.5% ROI over 8,700 bets, so it has been slightly below par at an overall level this season. This is disappointing in its own right but it’s safe to say it hasn’t been the best season so far and it’s reflected in the results we’re seeing here.

If we now look at my official proofed results which are meant to be a representation of the profits that can be achieved to potential subscribers (as proofed to the SBC), then this is simply the 6 combined established systems.  The SBC have been proofing these systems since inception and have been reporting back to their members on the performance.  This season, after a decent start in September, the systems have really struggled thereafter.

As you can see in the table, the established combined systems are showing a profit of 5pts from 588 bets. This is very disappointing and ultimately, doesn’t pain the service in the best light but importantly, I think it’s a fair indication of the season to date. The majority of subscribers follow these 6 systems and therefore, these results should represent the performance of the service so far this season.

Interestingly, if you break down the results by system, 6-21 has 19.4pts profit, 6-22 has 2.4pts and the other 4 systems are losing money so far this season. The 3 worst performing systems are 7-22, 8-21 and 8-22 and combined, these are showing a loss of 16.1pts from 124 bets.

I know from my regular exchanges with subscribers that a few decided to drop 6-21 this season since this system contained a lot of so called weaker bets and over the first two seasons, these bets were breaking even at best. On 6-21, it was the best bets that appeared on 8-21 and 8-22 that was making all the profits over the first two seasons and therefore, removing 6-21 from the portfolio of systems didn’t seem like that a bad idea.

As you can see, the strong bets on 8-21 and 8-22 are doing rubbish this season and must be pulling down the performance of 6-21. Therefore, it is the weaker bets on 6-21 that are having a good season this season.

I know there are a number of subscribers who are only following system 6-21 as this was the system which was looking to create the highest ROC this season from these systems although it would never achieve the highest ROI (all things being equal with the combined systems!).  So far, this strategy hasn’t been too bad and although the ROI of 6.9% isn’t great, it is still a profit of 19.4pts. The recommended bank is 45pts this season, and therefore, this system has generated a ROC of 43% even though most of the best bets have been losing most of the season!

On the other hand, I know a number of subscribers are only following 7-22 and 8-22 this season as these traditionally provide the strongest bets and so far this system, these bets haven’t been winning. 7-22 has a betting bank of 35pts and is 8.4pts down, so 24% of the betting bank is gone at the moment.

It’s hard to know what has happened to the better bets this season on these ratings. System 7 and system 8 are to blame and they’ve taken a profitable set of ratings on system 6 and narrowed down the bets badly. System 6 hasn’t had the best season but even so, 7 and 8 should be doing better.

System 21 is showing a profit of 13.9% and a profit of 41.4pts this season. This is stunning for a base set of ratings and if I look quickly at the season target, the target was 30pts profit this season at an ROI of 6%! It is miles ahead of target and it is easily the best set of ratings so far this season. Historically though, the results have never been as good as this on system 21 and therefore, it is overachieving so far this season.

On the same note though, system 22 has taken these amazing bets and narrowed them down very badly. The highest value bets haven’t been winning and ultimately, this has meant that system 22 is struggling. We know 22 is a better system than 21 based on the first two seasons, so it is amazing that when system 21 has a stunning first half of the season, system 22 has struggled.

I think overall, the established systems have had a poor first half of the season although there are a couple of gems in amongst the poor results so far. System 21 has been a shining light and 6-21 isn’t doing too badly. On the other hand, 7,8,23 and therefore 7-21 thru to 8-22 have really been struggling.

I have hinted at this in previous posts and emails but never really expanded too much on this up until now. I suspect this will be the last season that we see these systems in their current guise. I’ve said this lots of times over the first couple of seasons but I strongly believe that every set of ratings has a shelve life. The first two algorithms were built on data from 2000/01 to 2007/8 and backtested on data from 2008/9 and 2009/10. They went live in 2010/11. I have NEVER touched these algorithms at all since as at the end of the day, why touch something that has never been broken?

However, I know there is a theory out there that says at the end of every season, you should recalibrate your algorithm, taking into account the data from the most recent season or at least, refitting the data to take into account more recent data. I have shied away from this so far as it basically means you are following a new system every season and therefore, you have to put an element of trust into the system builder that they don’t backfit a system or build a new system with new data that was worse than the previous system but I’m confident I’ve enough experience now to ensure if I rebuild the first two algorithms taking into account more data, I won’t fuc* it up!

I know of another football service monitored by the SBC that updates its algorithm every year and basically, recalibrates the system (or systems) every season. I also know another ratings service that has kept the same ratings since 2006 and has added continually to the ratings by adding additional rating factors and ultimately,  they just create new systems every season, whilst keeping track of the previous systems too.

Therefore, there is a precedent in the market for doing something and as I’ve said since day one at this game, the moment you standstill, your edge can vanish. Clearly, the edge on the first two algorithms hasn’t vanished and it is still there this season but the strength of the ratings has really been called into question this season so far. The strongest bets for the first two season aren’t the strongest bets now on the first two algorithms and therefore, questions must be asked I think and I think action needs to be taken.

I think the other thing to consider is that the new ratings this season (systems 41-42 although I can see the performance of 43 too) are behaving exactly like systems 6-8 and 21-22 in the first two seasons. I’d even go as far to say that systems 31-33 are behaving like the first two algorithms did but this is slightly complicated by the fact 31-33 didn’t have the best of seasons last season in their first season!

Basically, I think the younger the system is, the better it is likely to perform as the algorithm is built on more recent data and therefore, it’s probably asking a bit much of systems 6-8 and 21-22 to be as good this season as the first two seasons when none of the results from the last two season are included in the algorithms.

Of course, this is a massive change to what my current subscribers are familiar with and therefore, I’m not sure how people would feel about following a set of combined systems next season that would effectively be unproven but then again, the New Systems from this season would be in their second season next season!

So, I think that leads us onto the New Systems this season.  214.7pts profit from 2,069 bets on all 11 systems or on the combined systems, it is a profit of 109.5pts from 811 bets.

An exceptional performance and I am as excited about these systems as I was about the established systems in their first season. I think these systems won’t be as good as the established systems in terms of ROI but the advantage these New systems have over the Est Systems is turnover. System 7-22 has only had 46 bets this season and will be lucky to have 100 bets this season. System 32-42 which is the equivalent New system has had 100 bets already this season.  That’s a big difference and ultimately, this means that the ROC that can be achieved with these systems could be much better than the Est systems.

For example, system 33-41 has generated a profit of 22.9pts this season from 59 bets. The target profit for this system was 9pts this season, so it’s more than doubled that at the halfway stage. No guarantee of course it will finish with anywhere near this profit but the point is, the turnover on these systems is exceptional and therefore, it does mean that we can probably achieve higher ROC figures than on the Est systems going forward.

The thing that separated the Established Systems from every other football system in the market was their ability to find the best bets (until this season!) and ultimately, these New Systems are showing this same ability. 33 is better than 32 and 32 is better than 31. 42 is better than 41.  33-41 and 33-42 are sitting with ROIs of 38.8% and 31.4% so far this season. Although they are overachieving, there is every chance they could finish up with an ROI of around 20%+ this season and that would make them directly comparable with 8-21 and 8-22 in their first season. 

As I said above though, turnover is key. 8-22 has only had 162 bets in 2.5 seasons of live results. 33-42 has had 53 bets so far this season. Hence, 33-42 looks like it could achieve comparable returns to 8-22 but over a higher sample of bets than 8-22 could ever manage. That would make 33-42 a better system than 8-22 in my eyes (and in most people’s eyes I suspect).

Of course, as I have been all saying all season with these New Systems, this is their first season and it’s important we don’t get too carried away but so far, so good.

This leads us onto the Misc Systems.  Going into this season, I didn’t know what to expect. I had high hopes for these systems last season but they were frankly, very poor. They were so poor that I actually stopped following these systems in my betting portfolio last season but at the season end, when I analysed the results, they weren’t great but they weren’t exactly a disaster either.

I took the plunge with them again this season with variable stakes on each of the systems and so far, it has went OK. Placing 1pt win on each of the systems would have generated a profit of 119.8pts from 1051 bets. A very nice return for a set of 6 football systems and they have repaid my faith in them a little. I think this is probably about as good as these systems are and I see them as solid 5%-10% footie systems and in my eyes, these sort of systems have their place in a betting portfolio. They won’t make you rich, they won’t break your bank, at times, they can have a lot of bets and not make much money but long-term, I think these systems have an edge and a decent one at that.

The betting bank of TOX is only 45pts for example and so far this season, the system has a profit of 32.7pts. A very decent return and although I know a few subscribers have taken the plunge with these systems this season as part of their portfolio, I wish more subscribers had taken a chance with these instead of following the Est Systems which appear to be the systems everyone has latched onto.

I’ll quickly touch on the Under/Over systems. Going into the season, I had no faith in either system after a very up and down season last season when they proofed in the SBC forum and ultimately, I thought if they broke even this season, they’d be doing OK! Well, a profit of 6.9pts from 168 bets this season is only slightly better than break-even. Interestingly, the target profit on the Over system is only 10pts for the season and it is at 9pts already. Unfortunately, the Under system has struggled all season and is currently 2.4pts down!

I don’t think my lack of faith in these Under/Over systems has wavered too much since the season began and ultimately, I don’t like the systems and I’m not sure they have much on an edge. However, a few have people have contacted me to say they are following the systems and although they make not be huge profit makers, they are good for turnover on Betfair for commission purposes and they don’t do much damage and the odds can easily be achieved on average although on each game, they can drift or steam wildly from when I give the bets out.

I think that only leaves us with the Euro Systems. Not sure how I try to analyse these as so far, the results have been crazy on a week to week basis. However, at a high level, 87.7pts from 1,007 bets and for the combined systems, a profit of 33.6pts from 416 bets. An ROI of 8.7% for all systems and an ROI of 8.1% for the combined systems.

Now, this all looks fine and well but when you look at the results by system, it isn’t exactly going to plan.

Firstly, the base ratings are proving very profitable. If I’m honest, the ratings are doing better than I thought they would do! The first algorithm has a profit of 13.7% from 181 bets and the second algorithm has a profit of 12.2% from 194 bets. Combined, system E1-E6 has a profit of 15.4% from 159 bets.  So, all fantastic results and it’s just a case of narrowing down these bets to get the best bets.

And herein lies the issue with the Euro systems so far this season. E7 has basically taken 89 of the 194 bets (so roughly 50% of the bets) and turned a 23.7pt profit into a 5.2pt loss. That’s very difficult to do but not impossible as we’re seeing.

Therefore, E1-E7, E2-E7 and E3-E7 are all struggling this season due to E7 fuc*ing things up. However, the systems not involving E7 are doing great and therefore, the other 3 combined systems are doing well and are making up for the mess E7 is making.

The other interesting thing about the Euro systems is the AH returns are awful! It seems to be the case the bets either win or lose which is strange as we’ve had lots of weeks where there have been plenty of draws but overall, covering the draw doesn’t appear to work. Not unless it involves E7 though as ultimately, the combined systems involving E7 are doing better when covering the draw!

Very strange but as I said up the top, the Euro systems are probably only 1/3 of the way through the season and in their first season, it is hard to draw too many conclusions at the moment.

Personally, I’ve been quite disappointed with the Euro systems myself as they have been so up and down but as always, you don’t really understand a system and how it performs until you follow it. At the end of the season, if the Euro systems all end up profitable, I won’t care too much about the ups and downs during the season but when you are living and breathing the footie as I’m doing this season, seeing so many ups and downs does take its toll on you.

I think that summarises the season to date for all the systems. If I’m being honest, it has been a disappointing season and I expect the range of returns for those following is massive. I communicate with a lot of subscribers via email regularly and I know some subscribers who have won 400pts (they follow every bet) and I know others who have lost 25%-35% of their betting bank this season already! It is a wide range of outcomes.

One thing I’ve said at the end of the first two seasons is that I’ve learnt a lot during the season that has helped me in future seasons and I suspect many subscribers will feel the same way at the end of this season.  With hindsight, I maybe could have directed more people towards taking a chance on systems which had a lower ROI but would potentially provide plenty of diversification in their portfolio but then again, I spent all last season telling people that TOX, STOY and STOZ would be good systems and they did shi* last season. It would have taken a bit of a punt from me to tell people to follow them this season!

I think the error we have all made this season is we have got too sucked into the Est Systems. Yes, they had the best results going into this season but following 5 or 6 of these systems and none of the other 35 systems meant all the eggs were in one basket and ultimately, the Est Systems have struggled this season and people are now looking at the results of other systems and wondering what would have happened if they’d diversified a little more.

I think diversification is the key here. I diversified a fair bit this season but I did the same for the last two seasons and at the end of each season, there were many others following the systems that had much better P&Ls than me. Ultimately, I looked on in envy seeing the fact they’d followed 5% of the bets I’d followed,5% of the time I’d put in and yet, ended up with a better £ profit and an ROI 3-4 times what I had achieved. My ROI this season for my portfolio is only slightly down on the first two seasons even though I’ve had people tell me this is a disaster of a season.

OK, I think that finishes the review for the season to date.

Here’s the table all the live results on each system to date.


I won’t go into the same level of detail as above but I think it’s worth touching on the live system results to date.

The Established Systems have an ROI of 8.5% across 6,976 bets and the combined systems have an ROI of 12% now across 2,580 bets. It is disappointing that the combined systems are seeing a reduction in ROI as the service has managed to keep an ROI of 15%+ since inception but 12% isn’t a disaster.

Looking at the pattern of the system results, it is very good. You can see clearly that the filtering works as you work up the systems, you can see that the combined systems improve on the single systems and lastly, you can see that the best systems are the higher combined systems.

I do think it’s interesting that even after a shocking season so far, System 7-22 and system 8-22 continue with an ROI of 17.7% and 19.4% respectively. I know these systems are taking a bit of stick and I know people are pulling their hair out this season following these systems but based on the live results to date, it’s hard to be too critical. 

Overall, the Established Systems continue to look OK in the long-term and although the short-term is disappointing (e.g. this season),  it only accounts for around 20% of the live bets and needs to be kept in some sort of perspective.

The only difference between the live results and the season results for the New systems are systems 31-33. These 3 systems were alive last season. The systems struggled a little last season (system 33 was the only loss making system in the TFA portfolio last season) but if you look at the results for a season and a half now, I think they are highly respectable. 31has an ROI of 4.2%, 32 has an ROI of 9% and 33 has an ROI of 9.7%. Not the best systems I’ve ever seen but then again, it’s not too bad given the level of turnover we’re seeing on these systems.

I think this shows that systems do bounce back. System 33 looked useless all season last season, made a loss and yet, this season, the very same system (nothing has changed!) is showing an ROI of 36%! Shows the impact that variance can have and also shows why you need to keep a long-term view at this game.

The Misc systems are really interesting as they have a decent sample of bets now since they were live last season. Overall, an ROI of 8.1% across 2,846 bets. Again, it’s not an amazing return but I like these systems. They were poor last season even though they made a small profit and they are very solid this season at a time when other systems are struggling. I think these systems offer something a bit different and a few of these systems should probably be part of everyone’s portfolio.

The only other systems which were live last season were the Under/Over systems. They struggled last season, didn’t get anywhere near the 10% ROI I was aiming for and this season isn’t much better. Overall now, an ROI of 5.9% from 641 bets which isn’t great. It is clear so far though that the Over system is the better system of the two and an ROI of 8.4% across 312 live bets is a decent return.

Overall, the service has now proofed 14,558 system bets which have generated a profit of 1,210pts. This is an ROI of 8.3%.

When I started out at this game, I hoped to achieve an ROI of 10% across all system bets and ultimately, this has proven a step too far. I then thought to myself that if the combined systems could achieve 10% overall, this would be a nice compromise and so far, this is holding up well. If I look at the all the combined system bets (including all 6 Misc Systems which are combined systems of some sort), then it’s a profit of 10.2% across 6,653 bets.  

Let’s hope the next 5 months see a decent performance from the systems and this will make up for the season’s performance to date. 

No comments:

Post a Comment