Print Story Colossal Fun with Betting Exchanges
Sport
By leviramsey (Tue Oct 02, 2007 at 03:09:36 PM EST) football, soccer, betting, betfair, quant (all tags)

Positions in the "to win the Premier League" market from last time, along with current prices (buy prices rounded up, sell prices rounded down; prices quoted from Betfair on the 0-100 scale at $10 per contract)

  • CHELSEA: short 127 contracts (avg price: 27.4): buy price is 12.2 => paper profit of $193.04
  • MAN UTD: short 147 contracts (avg price: 26.6): buy price is 36.8 => paper loss of $149.94
  • ARSENAL: long 33 contracts (avg price: 16.4): sell price is 28.9 => paper profit of $41.25
  • NEWCASTLE: long 50 contracts (avg price: 0.6): sell price is 0.2 => paper loss of $2.00
  • VILLA: long 123 contracts (avg price: 0.3): sell price is 0.1 => paper loss of $2.46
  • BLACKBURN: long 110 contracts (avg price: 0.3): sell price is 0.1 => paper loss of $2.20
  • LIVERPOOL: short 95 contracts (avg price: 21.9): buy price is 20.5 => paper profit of $13.30
  • MAN CITY: long 39 contracts (avg price: 0.6): sell price is 0.9 => paper profit of $1.17
  • WEST HAM: long 76 contracts (avg price: 0.2): sell price is 0.1 => paper loss of $0.76

Net paper profit: $91.40



Not to make this a series about the troubles at Stamford Bridge, but the fact is that Chelsea's fall is the biggest story of the season thus far. The home draw to Fulham, combined with the rest of the Country Club division of the Premier League posting away wins led to a 30%-plus drop in Chelsea's price, along with increases in the prices of the other three (which is good on only 1 of 3 counts).

The model's belief in Manyoo increased, to the point where Yoonited are now second-favorite for the title (though Arsenal are overwhelming favorites with an over 70% projected chance of winning). Man City and Liverpool round out the top 4, with Villa, Blackburn, and Everton projecting for the [potential] UEFA Cup places.

SELL 24 MAN UTD at 36.2. While the Second Club in the Second City looks (perhaps surprisingly) to stay up, a club of United's stature should be able to do better than win 1-0. Still, the difference between the price and intrinsic value has slightly narrowed, and thus the reduced sale.

BUY 12 ARSENAL at 29.5. Some trepidation in my heart, as Arsenal might eventually have to play a league game outside of the M25. Still, this is a heartless pursuit for a heartless bastard, or something like that.

SELL 43 LIVERPOOL at 20.0. Liverpool are a difficult club to figure, perhaps because of Benitez's squad rotations.

SELL 56 CHELSEA at 11.9. Even though Chelsea's price plummets weekly, and goes through my hedge-out point, I still have to load up.

BUY 13 MAN CITY at 1.1. Not as confident as I've been in the past, but still worth a play.

BUY 36 BLACKBURN at 0.2. False indications on Blackburn from earlier, but there's still enough potential for an upward move from here.

BUY 36 VILLA at 0.2. While it was breathtaking seeing Villa choke at White Hart Lane, the fact is that away draws (against a club that will improve its form) aren't bad things.

In total, that's $1,028.75 risked, and the positions now stand at:

  • CHELSEA: short 183 contracts with $1,415.24 risked
  • MAN UTD: short 171 contracts with $1,231.36 risked
  • ARSENAL: long 45 contracts with $89.46 risked
  • NEWCASTLE: long 50 contracts with $2.67 risked
  • VILLA: long 159 contracts with $3.43 risked
  • BLACKBURN: long 146 contracts with $3.44 risked
  • LIVERPOOL: long 138 contracts with $1,085.50 risked
  • MAN CITY: long 52 contracts with $3.77 risked
  • WEST HAM: long 76 contracts with $1.52 risked
  • Cash balance: $163.61

If $CLUB wins the league:

  • Chelsea: loss of $746.39
  • Man Utd: loss of $626.39
  • Arsenal: gain of $1,533.61
  • Newcastle: gain of $1,583.61
  • Villa: gain of $2,673.61
  • Blackburn: gain of $2,543.61
  • Liverpool: loss of $296.39
  • Man City: gain of $1,603.61
  • West Ham: gain of $1,843.61
  • Anyone not listed: gain of $1,083.61
< And so much more | BBC White season: 'Rivers of Blood' >
Colossal Fun with Betting Exchanges | 2 comments (2 topical, 0 hidden) | Trackback
Question by anonimouse (4.00 / 1) #1 Wed Oct 03, 2007 at 03:58:22 AM EST
I would guess that club prices drop pretty fast in the "to win the League" after the first 5-10 games of the season. I'd thought you would've been betting in the "positions" market instead.

Girls come and go but a mortgage is for 25 years -- JtL
I would probably expect that rough phenomenon by leviramsey (2.00 / 0) #2 Wed Oct 03, 2007 at 08:11:06 AM EST

Though it's not like I followed this market in any previous year. Obviously, for every club not named Arsenal (especially, I would guess, since Arsenal has a game in hand over all but 5 clubs), there is a question of whether they run out of games before they catch the good run of form.

The valuation algorithm takes the number of games remaining into account in a couple of places. First, the 19 clubs not in the lead are rated by tragic number: essentially the sum of points that they must drop and that the leader must win in order to insure that they cannot finish above the leader (for the leader, the number is defined as the sum of points dropped by Arsenal and won by the second place club that would dethrone the leader); there's also a small twiddle to account for goal differences.

The formula is (38-games_played)*3+current_points-leaders_points. While a table constructed on this basis essentially overvalues games in hand (it assumes 3 points for the games in hand), it is still based on the degree to which a club is still mathematically alive.

The league table on that metric is as follows:

  • Arsenal: 98
  • Liverpool: 87
  • Man Utd: 85
  • Man City: 84
  • Blackburn: 83
  • Villa: 82
  • Newcastle: 82
  • Everton: 81
  • WHam: 81
  • Portsmouth: 80
  • Chelsea: 80
  • Wigan: 76
  • Birmingham: 76
  • Boro: 76
  • Sunderland: 76
  • Fulham: 75
  • Reading: 75
  • Spurs: 74
  • Bolton: 73
  • Derby: 73

To this tragic number is added a projected final points total, based on an algorithm that predicts the remaining games based on form to date and recently this season. A constant equal to the number of table points the leaders have is deducted from this total, a twiddle factor between zero and one is multiplied in (based on the number of indicators of mathematical elimination... one, for instance, occurs when the projected points total for a club is less than the number attained by the leaders; at the moment, every club's factor is 1) and the absolute value of the resulting total is raised to a computed power (with some special handling for the potentially negative case). The intrinsic value is then the club's share of the total of raised totals.


--
Could I be the next Lee Abrams?
[ Parent ]
Colossal Fun with Betting Exchanges | 2 comments (2 topical, 0 hidden) | Trackback