Automated Betting

Although automated betting predates betting exchanges (some Tote and Pari-Mutuel providers allow their customers to submit a batch of bets), this betting style has proliferated since the advent of betting exchange APIs.

Betting exchanges allow both lay and back bets to be placed and charge on net winnings rather than every bet placed (with the exception of Matchbook). This gives their customers the opportunity to trade – placing many bets to establish a position. With the mechanical, systematic approach that underpins trading comes the following question: can I automate this?

In 2004, Betfair released the first version of their API, which enabled trades to be automated. BETDAQ followed suit with the release of their API. The Matchbook API is not currently available for general consumption.

Even if automated betting is of no interest to you, the chances are that you will have encountered it: some of the bets that you placed on an exchange will have been matched by a betting bot, implementing a trader’s strategy. Betting exchanges are compelled by regulatory bodies such as as the UK’s Gambling Commission to alert customers about betting bot activity. For example, the Betfair Terms and Conditions contain this text: “some customers make use of programs designed to automatically place bets within certain parameters set by them…(“bots”). These bots may be active in any or all Markets at any time and you should not assume that you can place bets on quiet Markets that will not be automatically matched simply because the Market otherwise appears quiet.”

If, on the other hand, automated betting is of interest to you, the fundamental point to bear in mind is that automation won’t turn a losing strategy into a winning one. You need to first satisfy yourself that your manual strategy delivers reproducible profits over time before you attempt to scale it by automation via an exchange API. Otherwise, the only benefit that automation will deliver is that you will lose money automatically rather than having to sit at a computer or be glued to your smartphone / tablet whilst losing. A bot is only as good as the strategy it implements remember.

Automation allows you to scale your strategy i.e. you can increase the volume and regularity of the betting activity that meets the criteria of the strategy. You therefore also need to make full use of the mechanisms the API provides for monitoring your strategy’s performance and the validity of its underlying assumptions.

Form Analysis

Automation can be used by those who study form to place bets that conform to their strategy. The parameters for such a strategy would be what constitutes value back or lay odds based on an analysis of the form. The API could then be used to automate the search for odds that meet these criteria or to request for better value prices – what would be realistic to ask for would be based on an analysis of the current market overround.

Cautious form students might chose to adopt a semi-automated approach, where they review the odds their custom program has found before using it to go ahead and place the bets. For example, you may know that the odds for a particular horse are out of line with its previous form because that horse has an injury that has not been well publicised. Your bot cannot know this. The semi-automated approach allows your feel for your chosen sport to be blended with the labour saving benefits that a bot brings.

Technical Analysis

Technical analysis or trading is a fully automated betting technique that automatically places bets, which meet the conditions coded into the bot that implements the strategy. A classic example of such an automated approach is where a bot scans markets to find an overround of less than 100% on the back side or greater than 100% on the lay side. If the bot backs or lays all selections (as appropriate) in such a market, it will lock in a profit.

The first thing to bear in mind with this example is that you will be competing against many other bots all implementing the same strategy and therefore searching for overround markets. One thing you may do in response to this is reduce the percentage overround your bot is looking for (because markets with the initial overround amount were snapped up by other bots), reducing your profit. Other bots will then detect this and reduce their desired overround, recreating the initial situation where your bot cannot find suitable markets to be on.

This is not to say you won’t make a profit with this strategy it’s just you may have to accept a smaller profit margin than you initially envisaged.

Remember though that to guarantee even a small profit the odds your bot sees need to still be available when it comes to place the bets. Additionally, you need to consider the effect that a non-runner may have on your profits. For example, you lay all the players in a tennis tournament to guarantee a profit. The assumption your strategy makes is that you will receive stakes for the bets you lay apart from ones placed on the tournament winner. However, if a fancied player withdraws you will still have to pay out on the winner at full odds (there’s no "reduction factor" applied to "material runners" in tennis markets), but you will not receive the stake for the withdrawn player, as this bet will be void.

Thus even a relatively simple automated strategy will have to incorporate business rules and exceptions. This is the real challenge posed by automated betting. By comparison, any learning curve associated with getting to grips with the technology and programing is relatively straightforward.