‘Quantitative data is information about quantities; that is, information that can be measured and written down with numbers.’ – Dictionary definition.
We’ve all heard the one about the Englishman Irishman and Scotsman, but have you heard the one about The Banker, Aussie Rules Football and a quantitative genius? This one is no joke though, it’s a story of how – in layman’s terms – a computer whizz developed a program to win betting on Aussie Rules, and here’s the twist, had never watched a game.
Applying financial principles to AFL punting
James was crunching data for a bank in Melbourne’s CBD, a position he had been in for ten years since leaving Uni. As a hobby with a few mates from work – for fun – they built a model to try and predict the outcome of AFL matches.
“We were looking to apply our financial quantitative skills to the sports world.”
“My mate was an Aussie Rules follower, so we went with the sport he knew – I had never watched a game. We built a computer model, fed in all the data and let it predict the result. Our advantage was that we came from the finance industry which has some very good techniques for analysing data – and at the time, some of it had never been applied to AFL before.”
Initially they went back and back-tested three years’ worth of data, feeding in basic stats such as disposals, marks amongst many others that James naturally is quite guarded to share. Checking the results, they were confident that the model was beating the market, as their back-test model was showing a healthy profit. So they started betting purely off the modelled results and sat back. This little hobby on average has made double digit ROIs per year putting James in the top 2% of pro Aussie rules punters out there!
Model without bias
“Picking AFL was random, but it was an advantage for me from the beginning as I had no bias, I just let the data tell me. Having never watched a game I had no preconceived notion that say Geelong had a good record against someone or that some teams travelled poorly, that as a fan for many years, you would have. It was all data based.” James highlights both the weather and home ground advantage as two key factors in a sport of many variables that bookies don’t always get right. “Some teams play a lot better at home, teams in Perth have a massive advantage with the time difference plus the fact not many away team fans are willing to travel that far – I really feel fans are a big factor in how umpiring plays out. Also, when North Melbourne play in Tasmania it has a home ground feel although obviously not as much as when they play at Etihad, while weather is linked to points scoring.”
Naturally decision time was around the corner and James, despite opposition from family and friends, turned his back on a high and regular income stream to turn pro. Leaving comfort and security for the unknown is always a tough moment and its only in hindsight you can realise certain facts, “It was the best decision I made, I shouldn’t have been so worried at the time looking back. I think the big factor in my decision to become a pro punter was knowing clearly what I DIDN’T want to do. And that was sitting working in a bank for the next 25 years.”
James admits that being older (30) and with the safety net of being in a strong financial position meant there was less risk – but at the same time made it harder to leave an already established career.
“I think now that if you’re young, have just finished Uni, you’re passionate about sports and numbers and are a quantitative person developing a system that you are confident can beat the market, consider making that jump and taking a year off.”
“The worst that can happen is that it doesn’t work; or it turns out you don’t enjoy it as much as you thought you would and you start a normal career a year later than you would have. Take a gap year. Be decisive and take advantage of key moments – punch the numbers and see if you can beat the market. If you can beat the market with simple models, then when you get more technical the only way is up, and you’ll learn a lot on the way.”
When James turned pro it was a quick learning curve as he decided to computer model other sports and as he freely admits, tried to do too much. “I thought I could beat all sports but some such as tennis were too efficient as the big syndicates had information that I would never be aware of. It was time consuming and my tennis model that I had spent three months on was just breaking square, so I refocused and developed a horse racing program which has now been going for six months and showing good numbers on UK and Aussie racing. When I turned pro I wanted to diversify – I treat it like a share portfolio, so my money is in different areas such as AFL, rugby, racing etc – a pie chart showing the proportions of my various investments.”
Higher scoring makes for better modelling
James makes an interesting point that AFL lends itself to computer modelling with its generally free flowing scoring, “NRL is harder to model the way I’m used to as there is stagnant scoring –while some other international sports markets are generally already pretty efficient as the big syndicates have data and information not available to the general public. 100% of James’ horse action is on BETFAIR while 50% of his AFL action is on the exchange. “I worked with BETFAIR’S VIP program and team resources to develop my horse race program, they are a tremendous help with data – such as horse SP’s in the last three years and so forth.”
With James’ AFL activity he has been successful betting on the total game points (over/under) however getting set is not easy, “It’s a smaller market really and there is not much liquidity here, unlike the lines market where you can get a much bigger bet on at least on game day. I will bet early in the week to take advantage of early corporate prices though once again it is hard to get set and sometimes it’s not worth the time trying to get on. I feel as long as liquidity keeps growing, BETFAIR is where you want to go, its betting in its purest form, its transparent there are buyers and sellers taking opposing views on match outcomes, its betting how it should be.”
James’ AFL strategy is pre-game match betting – no in-play betting (“it’s not my expertise”), no season betting (“not going to lock my money up for six months”). Coach’s talk in cliché’s about taking one game at a time, well it’s what James does – literally. The computer program spits out the prediction and James bets an amount according how much value he sees in the game. “Yes, it’s the same staking system for every game whether it’s the bottom two teams playing in round 11 or the Grand Final. If the computer makes the line -20.5 and I can get say -5.5 then of course I will stake more but its all driven by the program plus the value I can get in the marketplace. The staking strategy is pre-determined.”
Many punters will back with their hearts, especially in Aussie Rules where the fans devotion to their team borders on religion, meaning there is a high level of mug – perhaps recreational is a nicer way of putting it – money in the market. This is an edge that James exploits. Having said that, James has noticed smarter money coming into the marketplace. “There are syndicates popping up in AFL; I think they have realised AFL has been relatively soft and have jumped in. I’ve noticed some very smart movements last season that definitely weren’t there in prior seasons. I feel though there are pro’s and con’s with the increase in smart money (especially as BETFAIR liquidity increases). My ROI may well get eroded however as more money comes in I will be able to get set for larger amounts. If my ROI reduced but I could double my turnover, I would be very happy.”
In what is great advice for any player, especially if you’re starting out, are James’ comments on variance.
“Reassess each month not each week or each round. If you go 0/7 your model might not be wrong, likewise if you go 7/0 you might not be right – variance is key so analysing each month is the way to go.”
“The biggest mistake a lot of punters make is that they actually act like a gambler. Don’t chase, don’t increase your stake. If you have won your last five bets don’t up your stake. The only time you should have a bigger bet is if there is bigger value or if your bankroll increases to a point that you can afford it. If you have lost your last five bet’s it’s the same strategy. Pretend each game is the first game of the season since how your bets have gone on previous matches is irrelevant.” For someone who had never watched a game and is now making a very good living out of it surely the TV gets turned on when the players run out. “Yes! I watch the games now, I do know the players and it gives me an extra edge – the model is more sophisticated than it was to begin with. I do think that the market over reacts to injuries – unless it’s a key superstar – a lot of clubs have the roster to cover. In general injuries are over-rated and it’s important to take advantage of the markets over reaction.”
Managing your bankroll is crucial
James believes bankroll management is more important than being the smartest punter in the room.
“The number one pitfall, for the majority of punters, is bankroll control. You can have the best program or be the best judge but if you up your stakes chasing losses or getting overconfident when winning, then you will surely go bust.”
“If you are starting out, get the bankroll process locked down, and if your model is strong everything will fall into place. Don’t be scared to partner up with someone and enjoy the challenge of winning.” James is confident in his model that even when a big syndicate is going the other way he is not put off. “On the day of the game when I am betting head to head and it is clear that a big syndicate is opposing the model, I will generally trust the model. With BETFAIR it is such an advantage with the transparency and being able to see where the money is loaded and which way the market is moving.”
James developed a computer model to successfully bet on Aussie Rules football without ever having watched a game. He left a highly paid job with a blue-chip company to begin life as a professional punter. You could make a movie of it. I wonder how many computer programmers with quantitative skills are ready to break through the traditional risk averse barriers ingrained in them to pursue a career in the world of professional sports betting – the answer may well be many more then we think – as I’m sure right now, smart punters are looking for them.