Saturday, October 23, 2010

Forensic "Accounting": Is This High-profile Betting Exchange Minding the Store?

Betfair Group PLC had a high profile IPO on Friday (yesterday, October 22nd 2010) in London with shares doing well in aftermarket trading. Betfair (and its closest competitor, Betdaq) are a different breed of online bookmakers: they are the so-called "betting exchanges" where anyone can be the bettor or the bookie, and the odds are determined by supply and demand. Further, one can place a bet at the prevailing market, or can place a limit order of sorts at higher (or lower) odds in the hope of getting it filled if the market moves. One can trade in and out of positions as many times as he wants to, and can act as a market-maker, quoting bid/ask at any time. Unfortunately, Betfair does not accept US participants for regulatory reasons, but one can still poke around the site.

For any game, Betfair gives not only the odds but also the liquidity offered at each level, as well as the total matched (traded) and unmatched (bid/offer) volume so one can find the more liquid games where the spreads between the back and lay are tight. To back is to bet on a certain outcome; to lay is to bet against that outcome=what the bookie does. Betfair collects a % from the winnings for the service of matching counterparties, and, unlike a traditional bookmaker, does not act as a counterparty in any transactions.
It should be noted that most of the volume is in Western Europe soccer, and since soccer is a low scoring game, the typical bet is a straight-outcome bet (moneyline) and not a spread bet. There are three outcomes, home win, draw or visitor win (denoted as 1x2; and the home team is listed first, unlike the US standard). The odds are quotes as coefficients, that is the multiple of your bet you can expect to have back. For example, a 1.25 means you get $1.25 total back for a $1 bet (25% ROI). This is the equivalent of a -400 moneyline (you get 500 total back for your 400, 25% ROI).

I decided to check where the liquidity is on Friday night. One can search either by unmatched bets (bid/offer volume) or matched bets (traded volume). A decent % of the bets listed are longer-term bets (i.e. England Premiership winner this season), with some of the upcoming games mixed in (i.e. England Premiership weekend games). So imagine my surprise when I looked at the following ranking for matched bets liquidity:

Most liquid games as of Fri 11 pm EST:
Chelsea v Wolves $1,032k matched /$4,053k unmatched
Spartak Niltchak-Tom $527k matched/$101k unmatched

Tottenham-Everton $501k matched/$1,360k unmatched
Real Madrid- Santander $428k matched/$3,430k unmatched
Stoke- Manchester United $286 matched/514k unmatched
Birmingham-Blackpool $265k matched/$190k unmatched

There are a few REALLY odd things about #2. It is from the Russian Championship. Neither team is one of the premier teams in the league by far. There is a 5-to-1 ratio of matched-to-unmatched volume, versus 1-to-8 for the Real M game or 1-to-4 for the Chelsea game. A bit odd, right? Even the Birmingham game ratio is 1.4-to-1. So the data is showing lots of volume traded, but shallow depth, unlike most other games that show depth much greater than traded volume.

So this peaked my curiosity: since when has the Russian championship become so popular? I went to see the volumes traded in all the other games and I was in for a surprise. There was one game (Spartak Moscow, a high profile team) with $217k matched, but the other were $1.6k, $1.1k, $0.85k, $11k, $12k, $2.2k for the Alania, Saturn, Amkar, CSKA, Anzhi and Rostov home games respectively. The odd game has 50 times the average volume traded for all other games but one! Now this is getting really interesting.

Betfair also shows how the coefficients (the odds, basically) have changed over time. Based on my observations, the odds are fairly stable over the few days leading to the game, barring any major unexpected changes with one of the teams. So imagine my surprise when I saw that the coefficient for the home win (where most of the volume is traded, 521k of 526k total) has been DROPPING like a rock, from 1.60 to 1.25! So, in other words, you're seeing the strange "price" dropping on unusual volume over a few days, while bids/offers are uncharacteristically sparse at any level.

By now I was suspicious. I decided to check on WSN to see what is going on with the regular bookmakers and the Betdaq. Here are the odds quoted for a home town win: 1.57, 1.50, 1.53, 1.55, 1.57, 1.50, 1.53, 1.30, 1.23, 1.22, 1.53, 1.55, 1.53, 1.55, 1.53, 1.24, 1.24 from bet365, Wm Hill, bodog, unibet, expekt, sportingbet, Ladbrokes, PaddyPower, 188Bet, VCBet, BlueSquare, interwetten, totesport, bet-at-home, betsson, betfair, betdaq respectively. This is a HIGHLY unusual dispersion in quoted coefficients. It looks like some of the bookies had caught on the falling price, while other had not. The coefficient falls if there is large volume coming for the particular outcome (the equivalent of moving the line on a football game).

So what is happening here? To me it looks like there may have been some trading back and forth to move up the liquidity rankings, and once "dumb" money started coming in, there was heavy buying of "home win", moving the coefficient down. Again, usually the coefficient does not move much while here it looks like there was a big disbalance: someone was hogging all "home win" bids. The small number of unmatched bids is showing that there wasn't really a functioning market but rather an absorption of whatever "dumb" money came in to take the other side of the "home win" bet.

Was there an informed insider making the bets? Quite possibly, E European championships are (allegedly) rife with rigged games, reflecting the high levels of corruption in general. Traditional bookmakers will not take high volume bets on odd games, BUT the exchange gets paid on volume. So things like these can happen much to the detriment of the small bettor. If you think HFT and stock exchanges, you are not far off. You are not far off either if you think manipulation, "tape painting", etc.

Oh, yeah, and the final result: 2-1 for the home team. Color me surprised. Be careful because you don't know what the other side knows about the trade, and the deal facilitators will not look out for you.

No comments: