Saturday, 13 September 2014

Value Galore

With one exception, the number of selections, per entry, per round, is somewhere around 6.5. While I was hoping that most entrants would be cognizant of the fact that administering each selection takes time, and voluntarily restrict their entries to a few where they had identified value, I received 94 entries from one person for this weekend (81 in one entry, 13 in the other).

Not only is this a huge number, but many of the selections were Lays at higher (odds-against) prices, i.e. matches with a high win probability, which means a large number of bets where I have to enter the price into my spreadsheet and note the profit. Not a big deal for the odd one or two lays, but it gets old fast ploughing through 40 plus.

This is a huge amount of work of course, and with 35 FTL entries I clearly can’t commit to this effort every week. Even at an average of 10 minutes per entry per week, that’s close to six hours in total. This is something I do in my spare time, but I do have a full-time job and would like to keep my wife, so six hours is already at the top end.

Since it's my own fault for not having specified an upper limit, I'll deal with it this time, but as self-appointed commissioner of the FTL league, I am going to have to exercise my executive commissioner powers and limit the number of selections per entry per round of matches.

I’ll say 25 for now, but this is negotiable to a point. Some people's selections, my own XX Draws for example, are variable in number, and it could happen that one week they have no selections, other weeks they have 30 (highly unlikely, but possible), so if the number exceeds 25 but is usually below that, I’ll make exceptions. And I did email the person in question, and he is quite understanding of this change. I'm just glad to have got away with just 94 - in his reply, he mentioned:
I actually removed about the same about as I sent you because the edge was too small
Good heavens - there's a lot of value floating around this weekend! The other point that this strategy has raised is that the formula for adjusting prices to accommodate lay bets isn't sufficient in some markets. The rather naïve assumption on my part was that lays would be mostly at odds-on and shorter odds-against, but the fixed formula needs to be changed so that higher lays (2.5+) are being made into a market that mirrors that for back bets.

The screenshot below shows an example over-round of 1.03 (which was the average for Pinnacle in the Conference last season, with a range of 1.018 to 1.038):
Under the previous formula a Pinnacle back price of 3.2 was adjusted to 3.25 and a winning lay made 0.44 points profit. The new formula adjusts that 3.2 to 3.395 and makes the potential profit 0.42 points. It's a change that should only make a small difference, but if there are markets with larger over-rounds, the current formula leaves a loophole that could be exploited.  

Again, I'll stick with the current formula for this week, but the new formula will be effective after the weekend.

As for this weekend, we have a few debutants in action. Ian Erskine (of Erskine Cup fame, but also known for his FTS Income service), has ten selections, Daily 25 has one, Football Investor has seven, and Gecko has five. I already mentioned that TFA has his first (draw) selection this weekend. Football Eite is idle this week, as is @ValueBankFooty - I was looking forward to another accumulator. 

Here are the Wisdom of the Crowd numbers, representing where entrants see value:
A lot of support for the Arsenal v Manchester City draw, and no love for West Bromwich Albion or Manchester United (nothing new there). Only one clear home selection (Crystal Palace, and I have them as 7.2% value (currently 2.24 versus my 2.09). 


Unknown said...

I did try to warn you about the lay prices!

I'm happy with people making as many selections as they like as surely the more instances of perceived value you have the higher the likelihood that your perception of value is wrong.

But then again i'm not doing the admin.

If I had a method that picked that many selections which were value I wouldn't be part of the FTL at all as I would spend all my time calculating trying to get bets on & applying the same method to as many leagues as possible.

I do think you should make it so that everybody has to submit their bets in a specified format (using football-data team names for example) so the admin is kept to a minimum.

TO said...

Hi Cassini,

The admin sounds like a nightmare. I run a horse racing tipster competition at High Class Equine, both paid and free entry, in a winner-takes-all prize and small win if a free entrant wins. Luckily we have just a small group of about 15 tipsters and one selection each Saturday, so manageable.However, it can be time consuming and if by chance a number of new entrants arrived I would quickly shut the door.

There are only so many hours in a day.

I can imagine the look on your face when 94 entries arrived.

You deserve a medal.

Kind regards,

BotTrader said...

You need to automate things especially if you use things like Football data's csv files for results etc. Shouldn't be the hardest thing in the world to set up a webpage that'd gather all matches etc

I used to use football data's files to do everything from the upcoming matches and results/odds.

Where/what do you currently use for the matches/results/odds at the moment I'll see if I can tweak my code to cobble something together

The Soccer Doctor said...

Valuebank footy is a total biff, as he forgot to email for the second week running - This week costing me a 10.6pt profit. That said, what goes around comes around as last week would have seen a 7.35pt loss, so not all bad but selective memory in forgetting last week and remembering this would have been a nice break - Reminders set everywhere now - and I am looking forward to a midweek thrash out and landing a nice little acca

fairfranco said...

Did he make a profit from his 91 selections?