Thursday, 3 May 2012

Ways to play system 7-22

I think one thing that will ultimately separately The Football Analyst service from other more traditional football tipping services in the long-term will be the ability to follow the systems using a range of different ways to suit your own risk aversion.

I know for a fact that system 7/22 is the system that is followed by most subscribers at the moment. It is being followed by many people who are operating a portfolio of systems and for the vast majority of people who are only following a single system, 7/22 is going to be that system.  Of course, I know others are following 6/21 or 7/21 or even just 21 but at the end of the day, everyone can follow the system or systems that suit their own risk aversion but also the number of bets that they feel comfortable with. 

Anyway, for this piece of analysis, let’s just concentrate on system 7/22.

I’ve touched on this before on the blog but next season, I want people to think more about Return on Capital and less about Return on Investment. Of course, the biggest selling point of my systems at the moment is the fact they can produce high ROIs over a large number of bets but as I referred to in this post here, it really is ROI for show and ROC for dough.

So, here’s an example based on system 7/22 only. I looked at comparing the returns of outright betting and DNB in that earlier post but let’s take this on another step now.

Obviously, with 4 ways to play the bets (Outright, AH0, AH0.25,AH0.50) and an infinite way to stake (could be anything from 0 to whatever you like), the intention of this isn’t to show you every possible way to play the bets.  However, I suspect that the more work you do yourself, the more you can maximise the returns.

This is a true story but last Summer, I was contacted by a new subscriber (who I now exchange emails with every couple of days!) who wanted me to test out a staking plan.  After exchanging emails, he settled on a staking plan which I thought was the strangest staking plan I’d ever seen! He’d split the Home and Away bets up on a number of systems and basically, he had a staking plan that went to 2 decimal places for each bet!  For example, system 21, 2.39pts on Homes, 1.13pts on Aways.  I’m not going to post his staking plan or his systems but here’s the whole point of this….

I asked the guy for some feedback for my new website (I’ve asked a small selection of subscribers for feedback which you’ll see when the site opens up for public viewing) but this particular guy has an ROC of nearly 200% this season!  In a way, I find it quite satisfying that someone who spent a lot of his own personal time last summer has ended up with an unbelievable return this season from his own staking plan and his own idea about what systems to follow.

Of course, I’m not saying everyone should do that or be annoyed that they’ve not tripled their bank this season (a 50%+ ROC is good in anyone’s eyes, so if you’ve made in excess of that, you’re doing well) but the point is, they don’t have to do this work themselves unless they really want to!  By reading what I write, exchanging the odd email with me, asking me to look at a sample portfolio and so on, everyone can get a portfolio of systems and a betting method to suit themselves.

Right, back to system 7-22 as I got distracted there!

Here’s a table which shows 9 ways to play system 7/22 based on the results to date. I’ve used the full history of results (backtested and live) but this is one system where the backtested results are much in line with the live results, so there is no bias towards previous seasons like there would be with some other systems that aren’t in line with backtested results.


As you can see, it is fascinating reading.  Simply backing 1pt on Homes and Aways to win outright is by far the highest ROI you can achieve on this system.  An ROI of 33%.  Not bad for a football system!

However, when you actually analyse it, if we use max drawdown * 2.5 as a proxy for the betting bank needed, then the ROC is only a whopping 607.2%. I say only but that basically means you’d have doubled your betting bank every season for the last 6 seasons on average! If you use compound staking of some sort, you’d be retired to the Bahamas now I think and reading this on your IPad!

Anyway, the point of the table is to show that there are other ways to make 7/22 pay.  I know Rowan across at TPI plays 2pts on Homes and 1pt on Aways on 7/22 as his preferred approach after we discussed it last Summer and yeah, this gives a better return on capital than simply backing every team outright.  He made a good decision this season as the ROI on this is better than backing outright this season!

The reason for this post is actually to do with Rowan as we were discussing his plans for next season and I commented that I thought most of the volatility on system 7/22 was being driven by the Aways, so he probably had to look towards using some form of AH betting on the Aways to reduce volatility.  Incidentally, I also commented that I thought 7/22 was no longer a system that could be followed in isolation due to the low number of bets it is producing but that’s not a topic for discussion here at the moment!

I think this table highlights the point I was making in the previous post about how we need to start looking at ROC more and less at ROI.  Yes, having a 33% ROI might be a great return but is it not better to have an ROI of 26% if you can use a much smaller bank and therefore, stake higher and achieve a much higher ROC? 

It’s these sort of questions that I think people need to consider for next season.  Of course, with so many systems to follow and so many ways to play the bets, I don’t expect anyone to achieve the optimal solution but due to the fact we all have different risk aversions and want a different number of bets, what is optimal for one person isn’t going to be optimal for another.

I hope people find some of this interesting.  With 19 out of 20 profitable systems this season for outright betting (20 out of 20 for AH0, AH0.25 and AH0.50), it’s been impossible to lose this season. However, as we look back on the season in a few weeks time, apart from wishing we’d staked more (always going to be the case when you have a good season), we need to think about what we can do to improve on this going forward.  Is it having more bets, less bets, variable stakes, using AH betting more, staking more on Homes, less on Aways and so on.

Although I’ll be doing my own review of the systems and I’ll probably give an update on how I did myself this season, it’s important others take a bit of time to think about their own performance when it comes to following the systems.

4 comments:

  1. Hi Graeme.

    As you know I've been spending a lot of time recently thinking about the best way of moving forward with a long term aim in mind. You, and anyone who reads my blog, knows that I think this will entail larger stakes being placed on football matches with bookmakers that will take them.

    With that in mind, as a general principle, I'm looking at increased stakes but lower volatility (not, I hasten to add, that 7/22 has been volatile this season - it hasn't and the suggestion of 2pts home bets and 1pt away was a masterstroke).

    Looking at that table, one thing smacks you square between the eyes. Amongst nine different ways of staking, the lowest roi achieved is 13%! What a starting point!

    Added to that, is the fact that the simple, straight 1pt/bet (ie. the method likely to have the highest level of volatility) has a maximum drawdown of just 16pts. Personally, I think 2.5 times the max drawdown for a bank on a system with six years of performance data is a little conservative, but even then, to achieve an ROC averaging just over 100% per season!?!

    The other thought that occurs to me is that you could easily progress through various staking systems as the bank grows (assuming a reinvestment of profits). The straight 2pt/1pt split has served me very well this season as mentioned, but as stakes get higher, a natural transition might be towards something like the methodolgy at the bottom of the table, where the drawdowns look relatively tiny but the ROC goes through the roof.

    Like you say, loads to think about.

    ReplyDelete
  2. Excellent post Graeme.

    I think you are spot on in suggesting that we individually think through how we want to use your systems, and look at the return on capital and levels of risk involved.

    Your systems are achieving some great results, but managing your own way through them is still key. I prefer to have about 10-20 selections that I will follow with a higher unit stake rather than 50-100 at a lower unit stake even if the 50-100 have achieved a higher % profit.

    It's definitely part of my pre-season plan to think through how I want to bet your systems next year having followed their performance this year.

    ReplyDelete
  3. Hi Rowan.

    Thanks for the comment. As usual, you make a lot of good points and I agree with you that depending on the stakes you’re using, you would ultimately adapt the staking plan you are using.

    Yeah, the drawdown comment is a fair one and I’m not best placed to discuss betting banks as my nature is always going to have you understaking! It’s a funny one as I personally overstake and always have done on these systems but when advising others (especially on this blog!), I get drawn towards a very conservative approach to betting bank. I guess it’s just that I don’t want to be responsible for anyone losing badly, so I try to ensure people are aware of worst case scenarios. As I’ve said all along, people should use a bank they are comfortable with. ;)

    No analysis is ever perfect and if you look at that table, you’ll see that the drawdown of 7.5pts appears 3 times. Quite simply, the Homes have never had a drawdown that’s been very bad in 6 seasons, so as long as you do something to reduce the variance of the Aways, you can stake more on the Homes and it won’t impact your drawdown.

    I suspect there is an argument for saying that you can include more Home bets in your portfolio as these tend to be very good bets and due to the less average odds (lower volatility), you could follow something like 7/22 and all 21 homes which in essence would mean you stake 2pts on all 7/22 homes but 1pt on all 21 homes with maybe 1pt AH0 on Aways on 7/22. I suspect you need to get the bet total up a little next season.

    I’ll do a post on it today but I have my reservations now about 7/22 as a standalone system. It may well be the best system on the planet and the one we’ve all got most faith in but with less than 100 bets a season (only 70 this season), this could break-even for the next 2 seasons and would still be the best system on the planet on overall returns. Definitely something to consider this Summer also.

    I think it’s good that we can all spend time discussing how to make best use of the systems this Summer rather than trying to find a profitable edge. Not a bad place to be in!

    Graeme

    ReplyDelete
  4. Hi Fizzer.

    Thanks again for the comment.

    Yeah, I think the thing that is ultimately my downfall in some people’s eyes (too many systems, too many bets) when it comes to attracting new members is actually my greatest asset in many other’s eyes. I knew when I developed my first rating algorithm that I was always going to end up creating another algorithm and therefore, if I’d stopped at systems 6,7,8 and did nothing else, I would still have a set of systems which are comparable with some of the best services in the market doing what I’m doing.

    However, I’ve taken that a step further and ultimately, it is actually the building of the second algorithm that has got me to where I am at the moment. If I’d just settled on the first algorithm and built more systems based on that, I’d probably end up with a lot more systems similar in return to 6-8 with ROIs of 5%-10%. Again, that’s not bad and before I started building any footie ratings, people were telling me that if I could get a system which had an ROI of around 5%-10%, I’d be doing great.

    Although I’m only 2 seasons into this project, apart from adding more leagues using the same methodology, I reckon I’ve taken the footie returns to the highest place you’ll see footie returns over a large sample of bets.

    My next task has to be to replicate this for European Leagues. If I can manage that (it’s a big IF as there are no guarantees in this game), then I think I’ll have achieved what I set out to do with this project within the 3 season timescale I set myself at the outset.

    My next task has to be getting this out to a wider audience. The footie betting market is a big place and although it’s great that the guys subscribing this season can meet or beat my proofed results, it’s not really sustainable to run a small service when the systems are the best in the market! I’d be as well stopping the service and working in a supermarket at night as I’d earn more per hour than I’ve done to date.

    I suspect you’ll enjoy the fact I’ll be releasing the historical system results to all members this Summer. Will give you plenty to analyse before next season.

    Cheers,

    Graeme

    ReplyDelete