Price charts and backtesting. Not financial markets with limited data.

Discussion in 'Data Sets and Feeds' started by plusEV, Apr 13, 2008.

  1. plusEV


    Hi, I discoverd this site a few months back when researching about about price charts.

    This will be a long post, a little background first:

    I've never bought or sold an equity in my life, and my only knowledge of day trading, trend trading, whatever, has come from this site, but I've taken in a great deal over the last few months.

    And I'm not here to learn how to be a supersonic trader. I'm interested in looking at games of chance with subjective propabilities in the same way that day traders look at markets.

    By subjective game of chance, I mean an event where the precise probability is impossible to predict. A casino makes it's money by knowing the precise probabilites and offering slightly under the odds. Bookmakers make their money by assessing probailities of events and offereing under the odds. This is what I mean by subjective.

    I've been working in betting markets for the last 10 years, it's all I've ever done, and for the last 5 have been an 'independant trader'. I put my money down whenever I think something is paying over the odds. The more I look into day trading, the more similarities I see on a fundamental level.

    My first job was working for a guy who had an automated system for betting into Aussie TAB horse pools. An old mathematician who would come in with a floppy disk every morning and spend 5 minutes the loading the data. What he'd done was find an inefficiency in certain bet offers on the tote that didn't agree with his assessed odds of each horse winning. I was to enter price data and execute the bets through an automated program (data entry and 3 clicks to me) as close as possible to race time.

    So I learned early on that it was possible to consistently win at anything where a price was determined by someone elses opinion. Two factors involved:

    1. Identify an edge.

    2. Execute.

    I agree with the line of simple is best. I was first able to identify an edge and execute it at a 'professional' level with online casino bonuses.

    The potential with casino bonuses with correct execution was massive. I was in early but lacked the resources to optimize before the edge was destroyed. Generous poker bonuses and lack of resources for players also provided a good edge five years ago.

    And now it's sportsbetting. For over 3 years I have been consistently betting edges in sports markets. My setups are fragile and I don't usually talk about sports betting. In public, people either ask for tips or assume you're lying. The sports betting forums are in their own little world and speak with esteem and expertise while making very little practical sense.

    I'm taking a bit of a leap here, and reaching out for some ideas to find new edges in sport. So I'm willing to talk a little about my approaches. Keeping in mind that there are probably 5 simple key points that if publicly posted, would destroy my current income. But I hope to transcend this level as I did the good ole casino and poker bonuses.

    Sports markets comprise of bookies, exchanges and players. Bookies and exchange seeds are the market makers. Every player is trying to beat them, and over 90% lose. The strong are fed by the weak, market makers and players alike.

    People who tell you x is paying y because the public are all over them, or the bookie has them short in anticipation of this public action, are out of touch. The market has changed real quick the last few years, and is driven by professional syndicates and advantage players, using various sophisticated methods to assess fair value. Main stream bookie firms are on the back foot, as all sorts, from maths geeks to team insiders will beat any value that the uninformed bias is likely to create into shape. And I know this from experience.

    I had some good mentorship at a bookie firm in Australia focusing on the then booming US sport market. Internet and IT had pretty much made anything anyone ever knew about pre 90's sports betting obsolete. My job was to monitor bets and move prices, an employed bookmaker.

    Now bookmaking experts will tell you the aim is to take equal action around the spread, balancing the book and ensuring profit. This is completely impracticle. The aim is to take bets at fair value minus the spread. A big bookmaking mistake is to ove the price of public bets, only to be beaten down with smart bets.

    Watching the smart money flow from the bookie's chair made me realize how much these syndicates influenced the markets, in a few different ways. I saw how bookmakers need extremely high volume to counter the smart money, or resort to drastic limiting measures.

    So what I do now is monitor the sports markets for occurences where there is a shift in 'fair value'. My approach is purely market based and I care little for sport. Who's playing is irrelevant, a team coming off a 3 day losing streak but playing their second consectutive home game on Firday means nothing. It's all in the market. Unless I have reason to beleive otherwise, the closing game time betting market is the most accurate indicator of the true price of an event. That last sentence is very important for anyone looking to take up sports betting.

    Any opportunity to exceed the closing price is to be pursued agressively, using Kelly staking assuming a 2% advantage (overlay), which concur with actual results over a significant sample size returning close to 2% on total amount bet. High variance is countered by high number of bets over time, making the long run faster.

    Expectation is assessed against the closing markets, quantified by long term results.

    The problem now is liquidity, not in the market itself, but much of my trading In the past has been on strong, common indicators where speed comes into play. The mass marketing arbitrage packages have hurt, severely reducing options, but I managed to keep growing by perfecting execution and adjusting the approach.

    The focus now is on betting into higher liquidty more often, else I will struggle to grow. I want to looke at what's happening before these big price moves, and also look to take more advantage of price movement that occurs gradually but one directionally.

    My intention is to create something to the affect of price / volume charts and backtesting methods for sport, nothing too full on, but ways of utilizing limited information to monitor price action.

    Right now I have access to a timestamped xml feed with price history for the last few years and a big number of events. The data is there to creat a time / price graph for every game, and I also have a theory for a good volume estimate relating to price movement. But I have no skills in this area, having little exerience with charts and backtesting. Examing historical price info could very useful, but the only information I can get is gibberish. This of course is a good thing. You also have to know where to look.