top of page

Odds Drift Filter

In this post I explain the odds drift filter applied to the automated lay betting services at Exponential Bet.

With lay betting (betting on a selection to lose) in particular the odds of the horse can determine your total bet exposure, ie if you lay a horse at odds of 7 for £1.00 your liability on the bet is £6.00 if betting using a fixed stake.

If you win the bet with the horse losing the race, assuming you're set at standard 5% commission (which I'd suggest you change to 2% immediately) you would make a profit of 0.95p.

If betting using fixed liability in comparison then your max bet will only be the amount you set, which in this example is £1.00, but your potential profit would return 0.16p for a total liability of 0.96p in this instance. 0.17p profit would tip you over the max amount at £1.02, but depending on bet timing this may happen to round the bet up or down.

I've covered fixed stake vs fixed liability at length in my blog posts so I won't go into the huge differences with these two bet styles in this post.

Additionally, you can find out more on fixed stake or liability betting on the lay betting service page, both at Racing Lays and Lay Dutching.

This is my most recent blog post giving detailed examples and comparisons here:

The Odds Drift Filter mentioned in the subject line, or the Anti Drift Filter I commonly refer to it (though habit) will check the max odds of a horse at the time the bet is placed, and if it's over the max odds, ie it's drifted in price from the time it was selected to be a qualifying bet, then it will skip that horse or race if it's the only selection in the event.

This filter is here to minimise the occasions horses are selected to lay with the automated strategies, but at the time the bet is scheduled to be placed within Cloud Bet Bot for members the price of the horse has drifted and is a lot higher than at the time of selection, which at that point it met criteria and was within the odds range set.

To break this down, with Racing Lays and Lay Dutching the horses Cloud Bet Bot is programmed to scan races to check if any bets qualify does this at a pre determined time before a race is due to start.

The times races actually start to when they are scheduled is another story completely, and I'll come to this shortly.

From the time the horse for Racing Lays, or horses in the case of Lay Dutching that has multiple bets per race arrives, the programmed time for this is set for example 1 or 2 minutes to post, ie race commencing time, at this time the prices may have gone up in price over the max setting.

They may have shifted above the maximum odds I'd like the bet to be placed, which for fixed stake could mean a liability on the bet that is not worth the risk, or with fixed liability a low profit that is again not worth the risk.

We have now added the Anti Drift Filter to assess the race at this betting moment, and if the horse meets criteria of max odds or less, then it's a bet.

If it's over the max odds at time it's skipped with no bet placed.

This reduces the amount of bets placed over the maximum level I've set for each lay betting service, but does not 100% eliminate bets being placed over this max odds setting.

Now for the theory vs reality gap.

In theory we should be able to tell Betfair we don't want any bets over a maximum odds level set, and that should be that.

If a bet attempt is made over the max odds level it would then be rejected and not placed.

This can be achieved using either a local betting bot, ie installed on your own machine or some online software that works to your own bet parameters. As in, created and set up by yourself running on your own PC or via a browser.

As an example it would lay a specific runner in a race at BSP odds, or even Exchange, and ignore the bet if over your max threshold set.

This means the software you're using is just acting for yourself and looking at the race at the time the bet is about to be placed for a specific set of filters, such as the position in the market at that moment according to odds for example, so it may be looking for the second favourite or third and so on according to your rules.

Some software allows for more filtering to be included, such as race type, class etc, and you can build up your strategies this way to create your own automated betting plans that only bet to the max odds within these rules at time of bet placement.

This is just one example of how it can be done using these platforms, but there are obviously many.

The main points here are that it's just operating your sole Betfair account and not hundreds of members at once, plus it's selecting the horse for your strategy based on the capabilities of the software you're using, and would not be choosing a horse at one time, then placing it at another pre race.

Another point to note as well is that you're relying on the software in most cases running locally on your PC, so it that dies a horrible death or the power goes off during racing you're stuffed.

Whereas with our set up it's based on Cloud Bet Bot alone being online and active, which you can login and access via any device, so you're not having to keep the PC on at home all day.

The online options obviously resolve this problem, but these come with their own issues and limitations. Such as regular updates and maintenance, which the downloadable ones also have. Which in comparison Cloud Bet Bot self updates and you'll never be asked to do anything technical.

Still though, with the online or downloaded options you have stricter parameters are possible, such as betting to these max odds settings at BSP, or time of bet.

Whatever the software you use it would be automating the bets for just you alone and filtering the odds at the time the bet is actually placed, this would give you the max odds filter you want set, and you're prepared to bet with for your own custom built strategies.

This would not be my strategies though, as mine run though Researcher and communicate with Cloud Bet Bot to select the horses to back or lay based on the criteria I've programmed in to Researcher and Nigel has then created to run within the bot.

This is a custom build set up allowing me to create strategies according to my own parameters, and not be limited by betting software available off the shelf.

I'm certainly not knocking off the shelf software as it can do things that Cloud Bet Bot cannot and I use Fairbot on a daily basis.

However, this type of software cannot operate multiple custom made betting strategies with total flexibility, and place the bets for hundreds of members online at once, fully automated, and giving members total control of staking options with different settings to select within each strategy.

Doing all this while accumulating results for each strategy live and gathering data that is stored for future development of the betting systems.

Some of the online platforms do provide excellent results and record keeping facilities, but not integrated to the level I use within Researcher.

Anyway, I digress a tad (far a change)..

The difference is, software you use for your own automated betting strategies are limited to the options you will find within the platform, which of course makes sense, and operate using just your single Betfair account either locally or online.

Cloud Bet Bot operates using my custom built settings and places bets for multiple users online.

This brings me back to the theory and reality gap, as in theory it would be great if we could set a max odds that every member is bound to with each bet, and as stated especially lay bets given the reasons explained above.

However, this is not possible given the strategies themselves are based on a live betting exchange, and with many bets placed at exactly the same time for each strategy we have an influence on the market.

Additionally, we have different bet types to accommodate such as direct to the exchange, BSP, fixed liability exchange or BSP, fixed stake direct to the exchange.

In order for all members to be matched there has to be room for manoeuvre, with an odds offset in place, and to allow for the factors out of our control, such as the time a race actually starts compared to the published race start time, which are very rarely the same.

These are temperamental animals after all, and getting them to post on time, or in the stalls when they aren't in the mood will affect a race starting time.

Sure Favs - Foundation version exploits this fact looking for the favourite to back at 1 minute past the post time if they meet criteria, but it happens so often that if criteria is actually met for a qualifying bet over 90% of these bets will be matched before the race jumps.

The point is, from the time a bet is selected to the time it is assessed not over our max odds setting, which for Racing Lays is 10 and Lay Dutching 25, even with the Anti Drift Filter things can change.

If the bet qualifies at the minute to post time I've set, ie 1 minute pre race, then it's a bet. If it's over max odds then it's not a bet at that moment in time.

From this time to the race jumping the odds can and will move, and this will affect direct to exchange bets or BSP.

If you're betting direct to the exchange for example and the bet just qualifies at odds of 10 for Racing Lays, this is the max odds to qualify for this strategy with the odds filter, so it goes on within the 1 minute run up to official race start.

This means that hundreds of member bets get requested to be placed approximately 10 seconds after the system chooses the horse, which allows enough time for odds to be collected, and then from here all the available money in the market on that horse will be gobbled up by our members.

If a horse is at very high odds then it's likely the money available will not be a lot, and it we're working on the fringes with some of these lay bets in races, possibly with a strong favourite, or with Lay Dutching at the bottom of the market, the money will go fast.

BTW we rarely have lapsed bets due to no liquidity, ie members not being matched, so getting the bets placed on our bets is not the problem. It will happen once in a blue moon, but of course BSP is the solution for this if that is a caveat you cannot accept. As with BSP you will always be matched once the horse has been selected and presented to Betfair.

In some cases with Sure Favs - Foundation exchange bets will go on ahead of BSP given the nature of it being 1 minute past post time, so a race can jump at that moment exactly with live bets being fired, but again this only affects a small portion of bets, and is highlighted on the page as a condition that is required to be accepted before betting with this strategy.

With the anti drift setting for the lay betting services however, it means from the point the bets are fired to market at the bet placement time they can be within or at the top end of the max odds settings, but due to member activity sometimes will drift over this level.

Nigel is looking to speed up the process of bet selection to bet placement to minimise the period of 10 seconds, which as explained is the time from when the horse is chosen to bets presented to Betfair.

This will not do anything to change the fact when our money goes into the market it will have an obvious impact, as it's a live betting exchange.

I report all my results to BSP odds which is the fair way for me to represent the services, as this is often a 50/50 split of better or worse outcomes on the odds compared to exchange betting.

The tools I use to generate the systems that are both live bet tested and historically checked, this marries up to these circumstances with the anti drift setting with complete precision.

If my tools are telling me it was a bet then it was a bet, even if it drifted over the max odds so there is no distortion between reality or testing.

The reality gap I mentioned above comes from running a strategy yourself locally compared to running online for hundreds of members, as this is a completely different operation required.

With the drift filter now in operation we've avoided a lot of unnecessary high odds bets so far in races that we just didn't need to be involved in, but a handful of bets have as explained tipped over the max setting due to the reasons outlined above.

As races will never go off exactly on time, and as these are fully automated betting strategies, factoring this in is just part of the process.

Given all published results are based on BSP it does mean that from bet selection, to placement with the anti drift at minutes to post odds, and then the final odds for the selection at BSP, we have a clear view and grounded foundation for both operating, developing, and reporting exact results over time with this setting. If it's a bet with the anti drift filter at minutes to post bet selection, then it's a bet for everyone, but like any Betfair live exchange betting it comes with some quirks.

Thanks for reading.


256 views0 comments

Recent Posts

See All


bottom of page