Interview with the Football Analyst – Part 2

9)      This year started off badly for you. Was it your worst run ever? Are you surprised that variance can be so big even over such a large amount of results and with such a proven edge?

In terms of a single month of results, September 2013 will always be etched in the memory of all TFA followers I suspect. The systems had suffered two very bad months before quite close together (Feb-11 and April-11) and therefore, in terms of drawdowns for the systems, Sep-13 wouldn’t be the highest drawdown on most systems. However, back then, I only had 11 systems running and therefore, any losses were spread across 11 systems. Sep-13 was a bit of a massacre with massive losses across all the UK systems and therefore, it felt much worse.

I think the other thing that is often forgotten is that the average level of stakes being used in season one is going to be much lower than in the 4th season after 3 seasons of good results and therefore, September has to go down as the worst month ever. For me personally, I lost a similar % of my betting bank as my previous worst run but the difference this season was my betting bank was 2.5 times more than when the previous run hit, so it was definitely more painful!

In terms of being surprised by variance, I’m never too surprised. I have had systems which have hit 8 winners in a row at an average odds of 3.00 and I’ve had the same system hit a losing streak of 20. Looking at it more broadly, I’ve seen runs on systems whereby they hit something like 15 winners in 18 bets and then 3 winners in the next 30 bets. 18 winners in 48 bets is probably a profit at the odds of the bets but at the end of the 48 bets, you’d be on your knees praying for the next winner!

I think the difficult thing this has season has been the results on the Established Systems. These systems have had 3 very good seasons, never came close to a losing season on any system and yet, this season, many of the systems are really struggling. This has surprised me a little but then again, with the way the results have fallen so far this season in the UK leagues, I’m not totally surprised. It hasn’t been an easy season generally.


10)      What’s the current state of play with your results this season to the end of January 2014?

Below are the results for the season to date. If you compare these with the historical results above, you will see how much more difficult this season has been for the systems.  Even so, the results aren’t quite as bad as some people are making out in my opinion. With a strong end to the season (touch wood!), most of the systems can have a profitable season and that looked impossible at the end of September!


Football Analyst Interview Image 4


11)      Are your selections price sensitive or are you picking out winners? Does it actually price up matches i.e. return odds?

This isn’t an easy question to answer as it’s a combination of both in reality. I think the first point I should make is that all my algorithms do generate a % probability of a Home/Away/Draw for every game I rate. Hence, I end up with 4 sets of prices for every game I rate in the UK and 2 sets of prices for every game in the Euro leagues. The Draw systems don’t have a value % generated as part of the algorithm as I couldn’t find value backing the draw in most games using my ratings, so value and odds play no part in the draw systems!

However, it’s not that straightforward as we have combined systems to think of too. If we take system 6. The first algorithm generates a minimum price for all bets on this system. Same story for system 21. The second algorithm generates its own prices and therefore, we can have a bet on system 6 and system 21 but both have different win probabilities and minimum odds. Sometimes, the difference can be substantial.

Let’s take last weekend and the Stoke v Man Utd game. Neither team appeared on systems 6 or 21. However, Stoke appeared on algorithm 3 and algorithm 4. Stoke had a minimum price of 5.28 on algorithm 3 (system 31) and they had a minimum price of 4.41 on algorithm 4 (system 41). At the odds of 7.27 on Thursday evening (Pinnacle), Stoke were 37.8% value and 64.8% value on each algorithm.

Obviously, the bet appears on system 31-41 too. However, I don’t give any value % for combined systems as ultimately, I couldn’t find a fair way to do it. System 31 is much better than system 41 long-term, so I tend to believe what System 31 says rather than System 41. Hence, averaging out value etc. is not a road I wanted to go down. Quite simply, none of the combined systems have minimum odds attached to them.

I guess the other question to answer is why did systems 6 and systems 21 not have Stoke as value? Well, in reality, both the first two algorithms did have Stoke as big value. However, the 5 home systems that sit behind algorithm one and the 5 home systems that sit behind algorithm two didn’t find Stoke as a bet. Hence, it didn’t appear on system 6 and system 21.

Value is only one rating factor used by my systems and this is why I believe my systems are better than just a simple rating algorithm as they are able to filter out the value bets that long-term, aren’t going to be winning enough to make it worthwhile. That’s not to say that they aren’t profitable but this game is all about maximising return on capital. Backing outsiders every week may be a profitable strategy long-term (it isn’t!) but if it was, it doesn’t mean it’s the best way to use your betting bank. My systems deliberately cut out a lot of high-priced value bets as they don’t win often enough even if there may be an edge there. It varies by algorithm though and algorithm 3 and 4 find many more high odds bets than the first two algorithms as there are fewer restrictions on the odds in the 10 underlying systems.


12)      What is your take on staking? What is advised for your bets?

My view on staking is that simple is best. All my system bets are 1pt level stakes and all bets are proofed to this staking plan.

In reality, all subscribers play their own staking plan and I do the same with my portfolio of bets. I tend to vary stakes myself by system and by Home/Away bets but nothing too drastic!


13)      I run a couple of systems myself. One prices up the premier league based on my % estimations and the other looks at shots data. Both have been moderately successful this season. How are your system bets produced and what factors are most important?

All of my rating algorithms are based on shots on goal data and shots on target data. I basically use 16 factors on each algorithm and I cover most of the common rating factors used by football models. Every factor is given an optimal weighting which is back solved using a GLM regression method in SAS based on the historical data I’ve used.

The actual rating algorithm isn’t my USP I suspect, the USP is what I do with the ratings and the system building. Everything is based around being able to filter the bets to achieve the best results and I think looking at my systems and the returns, you can see this is what separates my results from others in this game.


14)      Your systems produce a lot of bets. Which systems do you follow and why? What type of bettor is TFA suited to?

This is another good question!

In reality, every member plays things differently and I’ve seen many people take my systems and come up with really original ways of following the bets and make better profits than I can. Likewise, I’ve seen people take my systems, come up with original ways to follow the systems and make a right hash of things and make a loss. The best example was 2 years ago when every UK system made a profit and yet, one member lost money by following a subset of bets in the systems.

I think one point I make crystal clear to followers of TFA is that the overall P&L they make from the systems is their responsibility. Obviously, it is easier to make a profit if every single system of the 52 systems live makes a profit but that’s not always going to be the case every season.

I’ve shown the results for the first 3 seasons by system earlier but in summary, 100% of systems made a profit in season one, 95% of systems made a profit in season two and 80% of systems made a profit in year three.   This season has been tougher so far and only 67% of the systems are in profit at the end of January.

Based on these results, it’s easy to understand why TFA has been heralded as the number one football ratings service for the last few years by The Secret Betting Club and the service has won numerous awards in the Tipster awards that the SBC members vote on every year.

In terms of how I use the systems, I use the same strategy every season and it’s partly why I have so many systems nowadays! I like to diversify my risk by spreading my stakes across as many algorithms, leagues and types of bets as possible and therefore, I always follow a well-diversified portfolio. I play a staking plan that optimises the profits of my portfolio historically and every season so far, I have added more systems to my portfolio, this creating more diversity and increasing my potential leverage.

I always build my portfolio just before the season kicks off as I always give up a bit of time every summer to help other members build portfolios in Excel that they want to follow. Hence, a member writes to me to say that they want to follow System 6-21, 31, STOZ, D1-D6 and E2-E7 and I pull together all these historical bets into a portfolio spreadsheet I’ve developed, show the P&L split down by season, month, league, Homes/Aways and look at drawdowns etc. based on the staking plan they suggest.

The advantage of building my portfolio last is that I’ve already spent hours and hours playing with the data, looking at staking plans, correlation of systems and so on and therefore, I always feel confident that when I start the season, I have a portfolio that is well thought out and gives me the best chance of maximising my profits whilst controlling my risk to some degree.

I’ve been very vocal about the need for diversification and the concerns I have with correlation and ultimately, I think a full diversified portfolio is the key to making TFA work long-term. That’s not to say it’s the only way to make a profit. Looking at the first 3 systems, you could have just followed one system and ultimately, managed to make a profit quite easily every season. Same idea for this season as with 2/3 of the systems in profit, if you randomly selected a system to follow, the odds are in your favour.

However, the beauty of TFA is the ability to really make your betting bank work for you and if done properly, you can achieve a very high ROC (return on capital) without having to take too many risks. Of course, for this to work, you need the systems to play ball and make a profit and this season, that’s been more difficult than the last 3 seasons!

If we imagine a portfolio of 1 system. You might need a 50pt bank to make it work based on average odds and historical drawdowns. It generates 250 bets a season and makes 10% ROI. This gives 25pts profit which against a 50pt bank, gives you a ROC of 50%.

250 bets a season is roughly 6 bets a week and it’s not exactly making the most of your betting bank if the other 44pts in your betting bank aren’t being used every week.

So, you can add a second system to the portfolio, with the same number of bets and the same ROI. It achieves 25pts profit against a 50pt bank.

However, if you combine the results of system 1 and 2, you may find that you only need a betting bank of 60pts based on historical results and drawdowns and therefore, you have generated a profit of 50pts against a bank of 60pts. A ROC of 83%.

Hence, by adding another system to your portfolio and thus reducing your risk (all your eggs aren’t in one basket now!), you have increased the ROC to 83% from 50%.

This principle is what TFA is built on and as I say, when it works, it can really work well but if used incorrectly (building highly correlated portfolios as an example), if a drawdown hits, it can wipe out your betting bank if you are leveraged too much.

Over the first 2 seasons, I achieved ROC’s of 100%+ each season and last season was spectacular for me with an ROC of over 200%. This season, I’m only slightly better than break-even at the moment on my TFA bets although I have lost nearly 50% of my betting bank twice this season during two very bad drawdowns for my portfolio. Thankfully, I’ve won it back twice!

TFA isn’t for everyone and ideally, it suits people who have some Excel skills, understand how portfolio betting works, aren’t afraid of hard work and don’t mind spending a few hours a week placing bets and recording results.

Of course, if anyone wants to join and just pick a system to follow and look for a small return on their betting bank every season, that’s OK too but if the system underperforms or if the market conditions are difficult over a season, there is a higher chance they will make a loss than someone following a well-diversified portfolio.


15)      Do you advise one bookmaker or several? Which bookmaker would you say is the best for football betting in general?

In terms of the prices quoted by me for the proofed results and the value calculations in the modelling, I either use a Pinnacle price or a general price as long as it’s quoted by a number of leading bookmakers. The service has been applauded by members and The Secret Betting Club for the fairness of the odds quoting. I am heavily restricted with a large number of UK bookmakers now and place over 70% of my bets with Pinnacle and I achieve around 99% of my quoted prices. Bear in mind my average stake is higher than the average stake for the service I expect and therefore, it’s not a bad figure for me. If you look around Twitter and various blogs, you will see members are able to achieve 99%-101% of my quoted odds.

In terms of me personally, as I mentioned above, Pinnacle is the main bookmaker I use. It’s amazing how quickly bookmaker accounts are closed if you’re taking prices on a Thursday evening for football matches that kick-off on a Saturday and generally, with the edge my systems have, I’m backing at higher prices than available at kick-off.

For season one, I didn’t even have a Pinnacle account which shows how much my betting patterns have changed. Of course, I didn’t have the betting bank I have now back then and therefore, I don’t think my accounts would have lasted as long as they did if I was playing the stakes I’m playing this season!

In terms of my other bookmakers, Ladbrokes take second place and around 10% of my bets are with them. I generally place my bets on a Saturday morning now which means I usually suffer from slightly lower odds but the liquidity is better with the bookmakers and Ladbrokes in particular seem quite happy to lay me some chunky bets on my teams at standout prices, so I use them when Pinnacle don’t offer a better price. The other bookies all account for 1% or 2% of my betting activity and most have restrictions in place regarding stake size or how much I can win on certain games.

The advantage of having the Euro bets these days is that the liquidity is much better and given how competitive Pinnacle are, nearly all my betting for the European bets in my portfolio is done with Pinnacle.


16)      Thanks for the interview. As a last question, what does the future hold for your own betting portfolio and the service in general?

I think the easy answer is more of the same.

I really enjoying following a football betting portfolio given I work full-time and therefore, I don’t see me changing too many things in the next 3-5 years. Hopefully as my profits continue to grow, my betting bank will continue to grow and with Pinnacle (and the option of potentially using Betdaq or Matchbook or even going back to Betfair), I don’t see too many issues with getting on the bets at the stakes I play now or even slightly larger stakes.

This season is the first season where I have added another football tipster’s systems to my portfolio and so far, it has worked well as the results from these systems have been better than my own so far!  I don’t see me adding too many other systems to my overall portfolio although I do have my eye on the Fink Tank ratings and I may well look to add some bets using these ratings to my own betting portfolio in the future.

In terms of the TFA service, I hope the results take a turn for the better in the near future or I may not have too many happy members at the end of the season! Beyond this season, I do have some ideas for the service (not surprisingly, it may involve another set of systems!) but I think the big change is probably going to be around a tiered approach to subscriptions.

At the moment, everyone pays a flat subscription fee and receives access to all 52 system’s bets every week. However, as I mentioned above, I do give up a bit of my time to help some people with building portfolios etc. and this is basically something I do free of charge even though it is very time-consuming. Of course, I’ve got other members who just do their own thing, never ask me for help and therefore, I’m not sure it’s fair that they pay the same price as those who I give a lot of help to.

In addition, if this season has shown me anything, it’s that some people are better off if I just supply them with a portfolio of bets to follow with a simple staking plan and that way, they don’t need to torture themselves by seeing the bets on all 52 systems each weekend and seeing what could have happened if they had followed different systems etc. This is something I’ve been pondering for a wee while now and seeing the mess some people have got themselves into this season by following badly constructed and heavily correlated portfolios, I think there is a demand for me to just supply a list of bets to follow each week. This would also allow me to open up the service to more members who maybe aren’t interested in following a fully diversified portfolio of TFA systems every weekend!

If I can promise you one thing, it would be that I will never standstill at this game and I’m always working on the next set of ratings, the next set of systems or the next way to find another edge on football betting.


1 Comment

1 Comment

  1. Pingback: Interview with the Football Analyst – Part 1 | Betting Tools

Leave a Reply

Your email address will not be published. Required fields are marked *

Current month ye@r day *