Sunday, 14 November 2010

Quality v Quantity

Anonymous, not his real name I suspect, recently suggested that Football Elite, with their proven track record of the past few seasons, could learn from him, and that "they are having nowhere near enough bets. I think the non-backing of odds-on shots has been discussed here before. It's nonsensical."

Well, my thoughts on the subject were that it was the comment was nonsensical. There are 49 games across the major leagues of Europe most weekends. Football Elite has found a subset of these matches that provide long-term profitability. The folly of suggesting that this subset should be expanded to include already disregarded matches from the initial set should be obvious. Anyway, I was pleased in a way that Matt took time out from his perusal of next week's fixtures to address the comment directly himself, although I am also slightly concerned that Matt is not spending the time trying to make me more money! Get back to work please Matt.

Here is Matt's reply in full:

Hi Anonymous, if that is your real name :-)

If I thought I could make a good return betting in 1000 games a season then of course I would do it. I dont think there are that many games that have those opportunities though.

The type of bets that I dropped 2 and a half years ago were all making losses or only very, very slight profits (or in the case of lays would have been more profitable if I had backed to win instead of taking insurance on the draw)

My personal opinion is that it's better to focus on the one area that obviously has the significant edge rather than pad it out with other less strong or even loss making bets.

Regarding odds-on betting, it's not a case of 2.0 yes, 1.99 no. I just found that over the years I couldn't make a profit on short priced favourites in top class football. Thats not me saying there isn't value for anyone else just that I am rubbish at finding it at that end of the market.

I'm guessing the reason why is there is more "mug" money on short priced faves in top class football than any other betting markets so odds in general are artificially lower.

The sort of side I focus on (my niche I guess) is home sides in games where there isn't a strong favourite so there arnt going to be odds-on bets by the very nature of the games I focus on.

Obviously though I wouldn't rule out a bet because it was 1.95 and not 2.0. It's just the principle of not backing short priced faves rather than not backing odds-ons.

At the end of the day its just my approach. You might be right I might be going wrong somewhere but it works for me which is all that matters really!
For followers of the US code of football, the not-so-highly-esteemed Rich Allen (aka SBP) sends me his selections for today's upcoming games.

Minnesota Vikings -1
Indianapolis Colts -6.5
Tampa Bay Buccaneers -6.5
Seattle Seahawks +3.5
Kansas City Chiefs -1
St. Louis Rams +6.5


Anonymous said...

Time after time you show you don't understand some fairly basic points of gambling.

If you truly don't understand why someone who apparently is constantly reaping a double digit ROI is not having enough bets then there's little hope. Maybe another reader would care to explain.

Joep said...

As a subscriber to FE I'll give it a shot. For starters, Matt does offer more bets than the recommended bets listed on the website, he also offers so called "short list" bets, on average about 5 a weekend, that are on average so far profitable, yet not as profitable as the main picks. I personally see them as value picks, yet second rate compared to the recommended picks.

Secondly, FE recommends 4% stakes (or a 25 point roll) for every recommended bet. A tipster that uses such an agressive betting strategy needs to have a rather large edge in his bets.

Say for instance the average odds of FE picks 2.5, according to Kelly the picks need to win at least 42.4% of the time, which equates to a little less than 2.36 average odds. This means that the payout needs to be over 10% of the true odds (in this example) to be able to use such an aggressive betting strategy. If one would incorporate weaker tips, say those where the payout is only 5% over the true odds, according to Kelly, no more than 2% of the total bankroll in the same scenario. In order to maintain the same absolute profits in this system, the number of picks should be more than doubled with all long term profitable picks, which is a hard thing to accomplish in my opinion.

So by increasing the number of bets provided, even when they are slightly profitable, due to the increased volatility, you're forced to reduce betsize and thereby losing ev instead of gaining it.

I know that FE historically offers more value than needed in order to use optimal Kelly using a 25 point system, but I'm sure that's a choice Matt made to build in some security into his system, which would also be incorporated if he would choose to include weaker tips. Also, I think most, if not all, long term successful tipsters only use a fraction of Kelly instead of full Kelly, so his choice here is also very reasonable.

I hope that story made some sense to you and that my math isn't off :).

Joep said...

Oh, I did err on the safe side by using 5% edge on all bets in the 2nd example, could get rather complicated due to 2 unknown variables otherwise and I'm not exactly a math expert.

Anonymous said...

I'm not quite following your explanation fully.

One point. If the staking plan is based on Kelly, then why is each bet the same size (in % terms).

Even if you aim for, say, 10% value for each bet (too high, just an example), your perceived edge on each bet won't be precisely 10 % so therefore stakes should vary if using Kelly.

Joep said...

Considering the ROI of FE is over 10%, that number is not too high. Also, I think a fraction of Kelly using an average edge you think you have is a decent way of shaping your betsizing system. Considering it is very hard to pinpoint the exact value of each bet, different staking sizes would make things unnecessarily complicated.

The point of that long point was to point out that including weaker tips would increase volatility which in turn would lead to lower betsizes, which in turn would most likely lead to lower profits (unless you can find a large number of bets that are also having a decent edge).

Anonymous said...

Joep. 10% is too high, which is the point I am trying to make. The very reason I am discussing this is because the ROI of FE is too high. They're clearly not maximising their edge and should be increasing turnover with lower value bets, to increase their P/L further.