Friday, 12 November 2010

Derby della Madonnina

Another weekend coming up, and a full slate of quality football. Looking at the strong draws, and there are only four so it won't take long, the selections are Real Zarazoga v Sevilla, Wigan Athletic v West Bromwich Albion, Bari v Parma and Stoke City v Liverpool. Other 'weaker' draw selections are Everton v Arsenal, Caen v Lille, Kaiserslautern v Stuttgart and Lorient v Paris St Germain.

Value home picks include Bologna (2.22) v Brescia, Cagliari (2.34) v Genoa, Sampdoria (2.0) v Chievo Verona, Koln (2.18) v Borussia Moenchengladbach, Wolfsburg (2.36) v Schalke '04, Brest (2.04) v Sochaux, Montpellier (2.22) v Toulouse and Auxerre (2.44) v Rennes.

The Milan derby is interesting. The basic ratings have Inter (2.74) as better by 1.3 goals, but after adjusting for form they are 'only' favoured by 1.06 goals. Inter and AC are first and second in the Serie A ratings but in the 'official' League table, AC Milan are three points and three places above Inter at the top. Inter's form is actually the worst in Serie A, whereas AC Milan's form is behind only Juventus and Udinese. AC Milan are at 2.94, with the draw at 3.3. The ratings make a back of Inter value, but this is one example where the value could be too good to be true.


mouldhouse said...

I have inter down as a strong back and also i like brest (boom boom) - good luck this weekend.

Do you have a spreadsheet or document with the strong draw bets on, or the home value picks? It's quite clear that there's a lot of quality in the picks by the fact you so often front-run football elite. Appreciate the "This is NOT a P&L blog" tagline at the top so fully aware of the possibilty of being told to **** off.

Cheers :)

John said...

I just wondered how often you change/tweak your ratings system? At first I thought it would be a good idea to build a basic rating system then build on it and let it grow but now I'm not so sure. At the end of the day any rating system is a measuring device and therefore changing it will give you strange results. You wouldn't start building your house one measurement then change it once you get half way through, so is it then a bad idea to start tweaking a ratings system once you have begun collecting data? More importantly, does it actually matter if a rating system is accurate? Is it not more important for it to be consistent and allow you to take a more precise measurement of what's happening. If you are consistently off the mark with a prediction system that's better than mixed bag of results because you kept moving the goalposts so to speak.

Could the most basic of ratings systems kept as-is from day one provide more details than a regularly tweaked one. Shouldn't we be looking at and studying the results compared to the ratings to try and work out 'why' rather than tweaking the system to match the past results.

Anonymous said...

Talking of Football Elite. If (and it's a big if as I'm purely going off their website stats) their ROI is as big as they say then they are absolutely 100% not having enough bets.

To quote from the website. "I took a more selective approach for the 09/10 season as I eliminated any areas that I perceived to be weak links. That meant no odds-on bets, no lays, no cup games, no over/under bets and a more cautious approach at the start and the end of the season. I think the selective tactics worked well and meant that we were left with just bets that were "premium" bets."

If Football Elite do genuinely know what they're doing and have an edge then the tactics mentioned above are quite horrific. And if they make such enormous errors in that regard then I wouldn't trust them one bit in any other regard.

And, looking a little harder at their site, my lack of trust seems to be well founded. I read literally the first few lines of their "example" and stumbled across this gem:

"It looks unlikely that A.Madrid will score more than once and with their shaky defence pretty likely they will concede 1 or 2. Espanyol look a touch of value at their current price"

Does anyone really think that's good reasoning and a good approach?

mouldhouse said...

F-E are proofed to plenty of places - the only way to find out ultimately is to stump up and see for yourself I suppose.

I assume this blog gets advertised on the dreaded betfair forum - a lot of the anonymous comments have a lot in common with the betfair forum in general!

Anonymous said...

Fair enough if they're proofed properly at correct prices. So given the IF is now taken as being the case, then the rest of my statement stands. i.e. they are having nowhere near enough bets. I think the non-backing of odds-on shots has been discussed here before. It's nonsensical.

mouldhouse said...

Re: Avoiding odds-ons - I concur. There is nothing that makes an odds on shot, by definition, a poor bet. It limits your yield somewhat - of course. I suppose F-E goes for the balance between a winning strategy/ratings/whatever and his desire to deliver the best ROI, since that will be the yardstick by which all services are ultimately judged in the ever-expanding tipster market. Personally I prefer 1000 bets, 200 units of profit, 10% ROI to 200 bets, 100 units of profit, 30% ROI. Some people wouldn't - those without a more professional mindset might not be able to handle the drawdowns that come with including less ROI-heavy subsets of bets.

But essentially its tough not to agree - nothing makes a 2.00 shot a great one and 1.98 shot a poor one.