Wednesday 7 January 2015

Monthly Review of December 2014

Going to keep this monthly review very short. I usually say ‘short’ and write a few thousand words but being honest, I can’t be bothered discussing a month where not a lot went right for the systems. The underlying results were very good generally and therefore, I should have made a small fortune in December. The fact the systems didn’t even make a profit is disappointing but as always, there are possibly some extenuating circumstances for the results.

I’ll quickly touch on it but the Draw strike rate was very low for the month of December (22.8%) for the underlying results in the UK and therefore, the ratings and systems should have been able to capitalise on this. Unfortunately, they didn’t manage to capitalise and some systems have been impacted massively from Draws. You get these little runs where Draws impact my results but this month is maybe one of the worst months I’ve ever seen if I’m honest.

Below is a table which shows the Draw strike rate for the month of December for all 28 UK systems:


Sometimes, I don’t need words to explain. Given the average odds we’re playing at here, each of these Draws loses the systems 1pt when it would have been a 2.5pt profit on average. A swing of 3.5pts on each game on average. Multiply this up by a lot of systems and a lot of bets appearing on a lot of systems and you get the jist of the month for the UK systems. A massive loss that wasn’t far away from a massive profit. The best AH returns I’ve ever posted for a month but not much good for me and everyone else playing outright betting I’m afraid.

If we take the UK Draw systems, with such a low strike rate for Draws generally this month, I would expect these systems to suffer a fairly large loss. The fact they didn’t (they’ll post a small profit) says a lot about these systems and in a season that has been pretty tough to find UK Draws in the lower leagues, my Draw systems are continuing to fight hard to make a profit. Not going to be a vintage season like last season for the UK Draw systems but I’d take any profit given the Underlying results so far for Draws.

This only leaves the Euro leagues. The December strike rate for Draws was an amazing 32.1%. Therefore, I would expect my Euro systems to really struggle and again, the fact they are going to post a break-even month later is testament to these Euro systems. Might have taken me two attempts but I think the Euro ratings may actually have a little bit of an edge which is pleasing after the disaster in the first season with these ratings.

So, with a strike rate of 32.1% in the month for all games if backing Draws, surely to God my Euro Draw systems would have posted a massive profit and recovered all of the season’s losses? Well, no. They’ve posted another mammoth loss and I think this could be the worst set of ratings and systems I’ve ever seen. I thought my first attempt at Euro ratings would never be beaten but these Draw ratings are taking the pis* now with how bad they are.  Not sure I see any hope and the only question for next season is whether I have a second attempt at Euro Draw systems or if I just bin the idea completely. At the moment, I don’t think I can face ever betting on Euro Draws again but I’ll see how the next few months go.

This just about sums up my feelings about the systems this month and constitutes the review of the month at a high level.

I also have lots of other annoyances at the moment if I’m honest. I’m fed up with quoting Pinnacle prices when they are below market average for lower league games and ultimately, it cost the systems over 150pts in the last week of December alone. The liquidity for Boxing Day was appalling when I was releasing the bets and therefore, given the fact I’ve tied myself into quoting Pinny prices only for the past 2 seasons, I was forced to miss bets that were terrible odds with Pinny (other tipsters had released before me, causing the price to crash at Pinny but much better odds were generally available with UK bookmakers and even other Asian bookmakers).

It was as clear as night follows day that the teams would rebound up to kick-off and low and behold, the rebounds happened and the teams won. The fact I was on a couple of other teams (were false value prices as we ended up being on steamers that I didn’t think would win!) in the same games just rubbed salt into the already gaping wounds and to say I felt awful after the Boxing Day massacre is an understatement. In total, I missed 6 bets and of course, all 6 won. Sod’s Law some would say but I would say it’s stupidity on my part to be quoting false Pinnacle prices that are clearly too low to be used for running my value ratings. This needs sorted.

Putting all this together, with the first 3 months of the season being pretty rubbish before December and you start to get an insight into where my mind was at on the 26th of December.

What has happened since? Well, the big system bets have stopped drawing and started winning, the ratings have hit a bit of form generally and I’ve had a good few Euro winners too (not Euro Draws, they just keep losing money!). Given it’s only the 7th of January, it’s amazing how quickly things can turn and therefore, this sort of explains why I’m not about to spend a few hours lambasting my systems for a rubbish December. It happened, we were unlucky with hitting too many draws, we move on.

Here’s the results for all the different sets of systems in December.

Est Systems

Here’s the results for algorithms 1 & 2:




Losses on both algorithms and the only plus point (it’s hardly a positive!) is that the losses were with the lower value bets. Anyone filtering the bets by value would have missed some of the carnage this month.

Here’s the results for the 11 systems:


I think all you can say is WOW when you look at the AH returns. The highest AH0.50 monthly returns since the systems went live 5 seasons ago and when you look at the Draw strike rates above, you can see why this has happened.

The massive positive is that both the rating algorithms were rubbish, system 7 was rubbish too and yet, these 3 systems account for nearly all of the loss. Amazing that some systems managed a profit given how bad the ratings were this month and when you think these systems which managed a profit were also heavily impacted by Draws this month, they have been very unlucky.

I could pick any system but I will take system 7-21 as an example. 29 bets this month. 9 winners which resulted in a loss of 6.2pts.  What happened with the other 20 bets?

Well, these 20 bets included 14 Draws and 6 losers. 14 Draws ffs. The odds for these bets were 4.30, 4.07, 3.93, 3.91, 3.50, 3.38, 3.32, 3.27, 3.16, 2.98, 2.90, 2.83, 2.71, 2.22.

The same pattern is repeated across most of these combined systems and quite simply, December was maybe one of the most frustrating months my systems and ratings have ever had. Given the first 3 months were 3 of the toughest months they’ve had, it’s very frustrating that this has happened off the back of the results earlier this season.

Anyway, the Est Systems lost 27.9pts from 468 games.

New Systems

Here’s the results for algorithms 3 & 4:




Amazingly, algorithm 3 managed to make a profit (I do love this system!) and as you can see, it somehow managed to avoid the carnage on the other algorithms of hitting too many Draws. I don’t think it’s any coincidence that if you look back at the first picture in this post, which system had the lowest draw strike rate this month by a long way?  Yip, system 31.

Algorithm 4 lost again (4th month in a row) and I don’t appear able to stop the slide on this algorithm. I did a fair bit of work on it in the Summer but it hasn’t had the desired effect and the algorithm doesn’t have any edge and is possibly worse than last season!

One point I should make is that the low value bets caused the loss on algorithm 4 and also reduced the profits on algorithm 3. I know it’s not really a positive but given so many following are filtering bets by value, I think it shows these value ratings are at least doing their job to a degree. The ratings haven’t been as good this season as previous seasons but we’re only halfway through the season.

Here’s the results for the New Systems:


At face-value, it looks like a nightmare with a loss of 84pts from 507 bets but the Draw has been a killer. System 31-41 had a disaster and only hit 16 winners from 66 bets. What happened to the other 50 bets?

Well, 25 of them were draws!  Not going to quote all the odds for the bets on this system that were draws but it included odds such as 6.62, 6.33, 4.77, 4.67 and so on. Basically, the system was a couple of winners away from break-even and a handful of winners away from an amazing month. Closer than a loss of 84pts looks.

Having a month in a season where the Draw is a pest is to be expected but to happen after the first few months of the season the systems have had is just irritating really. Things can only get better!

Misc Systems

Here’s the results from the 6 systems:


Ouch!  Same story really with the AH returns being much better due to the impact of the Draws.  Nothing more to add.

A loss of 63.5pts from 290 bets.

Draw Systems

Here’s the results from the 11 systems:


A strange set of results really with the first Draw algorithm losing 3.3pts from 55 bets and the 2nd draw algorithm making a profit of 12.3pts from 67 bets.

This is the 2nd best performance in a month by the second algorithm since the ratings went live. Given it has happened in one of the toughest months for Draws since the ratings went live, it’s a very strange result.

I think overall, it’s an achievement for these systems to make a profit this month if I’m honest and therefore, I have to be happy. I keep waiting for it getting a little easier for Draw bettors in the UK leagues but so far, it hasn’t happened yet!

Overall, a profit of 10.9pts from 323 bets.

Euro Systems

Here’s the results:


A loss of 0.1pts from 272 bets. Same story as with the UK systems and the AH returns were better as a result of too many Draws.

Euro Draw Systems

Here’s the results:


A loss of 76.9pts from 291 bets. Given the underlying strike rate was 31% in the month, to only hit a strike rate of 21.3% takes some doing I’m afraid. They keep missing the winners and backing the losers, not really a recipe for a successful set of systems!

Overall, a loss of 241.5pts from 2,151 bets.

Never quoted this before on a monthly review but the returns from AH0.5 betting on all bets this month was 163.7pts! If there was ever a way to illustrate the impact of the draw on the monthly returns, this is it…..

Roll on 2015!

2 comments:

  1. The odds conundrum is certainly a difficult one for you Graeme. Firstly I think it would be fine to use one set of odds to allow the systems to determine whether it is a bet or not and report a separate set off odds results based on a varied sample of members and the odds they achieved in a variety of stake sizes at a variety of times all anonymously. This is something like the proofing benchmark I believe Australian Steve suggested.

    My suggestion for the odds calculation for determining the system bets would be to use a modified average odds from Oddsportal. I have my Oddsportal to show at the top level average odds rather than maximum odds. For most football games this will produce an book of about 106%. I would then scale the odds to a 103% book which represents an achievable book for most punters to bet into even with several limited accounts.

    So there you have it a modified theoretical set of odds for determining whether a bet is viable or not and then an actual field test set of results on the side for further confirmation that the theoretical model is producing achievable results.

    ReplyDelete
  2. Hi Matthew.

    Thanks for the comment. You make some good points and in principle, I agree with your suggestions but in reality, it is not practical for a service such as TFA. It actually could work for a service where everyone follows the same bets every time and I can see why having a sample of members supplying the average odds is a great idea but if I went down this route, I’d need to give up work to determine the proofing results for the service! :)

    The other issue is that the TFA value ratings are used by more and more now to determine what bets people have and what staking plan people play and therefore, any value ratings are dependent on the price quoted by me or price achieved by the subscriber. We’d have a situation where a bet which is deemed >5% value at release is not then a >5% value bet for the proofing results if I then use different odds for proofing and quite simply, a service such as TFA can’t have two sets of odds such as you describe. It would be a nightmare.

    The idea of an average odds is something I’m a fan of as long as it’s an average of a decent set of accounts. For the last two seasons, I always quote the Pinny price even though for some bets, the price is the lowest in the market as other tipsters have released before me and smashed up the Pinny price. However, for other bets, the Pinny price can be competitive and therefore, I figured by always quoting the Pinny price, I was using something that was fair and achievable long-term.

    I get fairly regular feedback on subscribers’ ability to match/exceed my quoted odds and I’ve even had good feedback from people in Australia and Hong Kong who are placing bets with massive delays and they do pretty well at getting my odds. Apart from the initial few weeks of the season when all of us are lumping on Conference teams in low liquidity markets, the feedback on odds quoted is very good for a service such as this and therefore, I personally don’t feel the need to change the odds policy from what I use at the moment.

    For the lucky ones with UK bookmaking accounts to use, they easily beat my odds by 2% or 3% and over a season, this is a substantial amount. Being honest, I’d never really monitored my own betting and how I do against what I quote for the susbcribers as last season, I was running at about 99.8% of my quoted odds, so there didn’t seem much point. This season, I have changed the way my bets are placed and the timing of the bets and as at 7th January (not updated my results a great weekend last weekend), my overall P&L is 26pts better off this season than the advised P&L if backing all the same teams 1pt level stakes. I’m beating the quoted price by 1.5% on average and this equates to 26pts! Admittedly, I place a lot of bets but even so, 26pts additional profit isn’t to be sniffed at and I think it has shown me that I’ve probably been missing out on a helluva lot of profits in previous season by being quite relaxed about when I place my bets. Definitely something I’ve learnt this season!

    Cheers,

    Graeme

    ReplyDelete