Following feedback on my previous post on my roulette history I thought I’d create a follow up; this time focused on my bets.
The purpose of the post is two-fold. Firstly I was interested in doing some analysis of just how many bets I’ve placed over the years but secondly and more importantly I wanted to address some of the questions that are often asked in Matched Betting communities.
This post is big, at over 2000 words and with lots of graphs and tables it does take a little while to load and if you’re not interested in statistics and numbers you may find this post a bit much.
Everything in this post is purely based on my historical bet records, unlike other posts none of this is mathematically proven and should all be taken for informational purposes only. Although my sample size is reasonable (at over 5000 bets) it’s by no means large enough to draw any definitive conclusions.
Due to the way I track my profit many bonuses are tracked as account adjustments rather than free bets so I’ve not been able to include them in the analysis.
This does impact the overall profit figures but hopefully the post is still useful.
How many bets have I placed?
To give a feel for the sample size this analysis is based on here’s some quick stats:
Number of bets – 7639 (I’ve actually placed slightly more but my first spreadsheet is lost to history
Number of back bets:
Free SNR’s: 511
Free SR’s: 87 (note that there’s another 858 account adjustments are excluded hence why this is so low)
Account adjustments: 858- these are a combination of bonuses that are account credits corrective figures due to errors and some bets that couldn’t be accounted any other way like spreadbets)
Lay bets: 2933
What type of bets have I placed?
Most of my bets are straight back and lays but I’ve also done a number of dutches, these are included in some of the graphs but have had to be excluded from others especially when I’ve analysed the impact of not laying as I don’t want 2 x ‘back bets’ from a dutch impacting the overall analysis.
What have I analysed?
I started writing this article with a simple analysis of profit over time, as it progressed I’ve added more and more questions into it until I’ve reached the point where I’ve analysed the following:
- Profit from all my bets?
- The impact of commission on profit levels?
- Are higher odds more profitable
- Stake Not Returned Bets – how profitable are they? should I have layed them?
- Profit from back bets only i.e how would profit have been impacted if I never layed any bets?
- Would I have profited without bonuses i.e. is taking our of line odds profitable long ter
Profit from my bets
Here we’ve got a graph mapping my profit from bookmakers. You can clearly see it’s been a relatively straight line of profit ever since I first started.
It’s worth noting that this is the only graph that includes the account adjustment bonuses, hence the larger profit number. This is the closest I’ve got to an overall profit figure from bookies.
Looking at the graph in more detail reveals a few points that look a little strange:
When we dive into these in more detail we see that:
This was my first big mistake Matched Betting, I found a huge arb (I now know this is a warning sign!) at a site called Turfez who are a Tote horse race betting site.
I placed a £200 back bet at odds of about 13 and layed it at a combination of 9.6 and 11. Little did I realise at the time that tote odds are only indicative and are actually based on the total prize pot being split.
My fairly sizeable bet ended up being a large proportion of the overall pot so my joy when the horse came in was short lived as ,y ‘massive arb’ only paid out at the equivalent of 1.6.
I can still remember the feeling when the realisation sunk in that I’d just lost nearly £900 due to not understanding tote betting. I now know that it is possible to profit from tote bonuses but you have to be very careful.
This is a graphing issue, I dutched Ladbrokes against SBO which resulted in one bet winning £1,250 while another lost £1,250, unfortunately they appear on the graph as a large loss and a large win separately.
Again another graphing issue this time winning £560 at totesport and losing £600 at 10bet while dutching my way through the 10bet WR.
Yet again another graphing issue due to dutching
My 2nd stupid big mistake matched betting, after Cheltenham this year I had a large balance in my Smarkets account.
I was rushing though a load of bets and clearly not paying enough attention as I accidentally layed Leicester in Crystal Palace v Leicester at odds of 50.
Although my lay stake was only £50 I still ended up with an exposure of over £2,000 and not the £75 I expected from odds of 2.52.
The worse part was It took me a few hours to realise my mistake…..
By the time I’d realised the error, the bet had been matched and I was left with a £1,400 loss and a sickening feeling in my stomach.
A warning to us all, even after matched betting for many years you can still make mistakes. If I’d checked my bet after placing it I could have cancelled the lay for a 2 figure loss!
Overall I guess you could say I’ve done fairly well to have ‘only’ 2 large mistakes over 7000 bets but there have been others, fortunately I’ve been able to spot and fix them before they seriously impacted my profit.
What about commission?
From a quick review of my lay bets it appears I’ve paid over £3,000 in commission even though my lay bets are split roughly 50:50 between Betfair / Betdaq at 4.8-5.0% commission and Smarkets at 1-2% commission.
What if I never layed any of my bets?
There’s a school of thought in matched betting communities that suggests we’d all be better off if we never layed a bet.
The main arguments for this centre around two key areas; the avoidance of commission and avoiding the need to place a bet (in this case a lay) at less than the true odds.
If we place a bet at odds less than the true likelihood of the outcome happening we lose money in the long run.
In order to simplify the analysis I removed all dutched bets and account corrections and simply looked at the back and lay bets, this reduced the sample size to a little over 5,000 bets as well as having a massive impact on the overall profit figure.
Graphing the back and lay profit shows nothing out of the ordinary. We can clearly see the tote mistake at Turfez, other than that it’s a straight forward case of gradual profits over time.
No real surprises, a steady profit and the very visible Turfez mistake
To make this graph more interesting we’ll now overlay the profit from the Back bets and the profit from the lay bets seperately.
When I first saw this graph I thought I’d made a mistake, I actually went back and checked my calculations a few times before I truly believed what I was seeing.
Overall my back and lay bets made about £6,000 (after ignoring account corrections) but until I did this graph I had no idea that was made up of about £28,000 of profit from the back bets and £22,000 loss from the lay bets.
If I’d never layed any of them I’d be over £22,000 better off
In fact once we’d completed the first 25 bets there was only 1 other period when we were better off due to having placed lay bets at the exchange (from bet 1486 to 1499). At all other times we’d have been better off not laying the bets at all.
What I’ve shown here is that I would have been much better off not laying any bets, this is a long way from definitive proof that it’s the best way to do matched betting and there are certainly other things to consider before we all stop laying our Matched Bets.
What about variance?
Having had the scary thought that I might have wasted over £20,000 laying bets I started thinking about some of the other benefits of laying them.
The first one being variance, the main benefit of laying bets is that it smooths out the variance and prevent the huge profit swings.
Lets remove all other data and just look at the profit from the back bets in isolation.
We can see from this that there is a fair amount of variance (as we’d expect), much of this caused by some disproportionately large bets. The first time profit drops significantly (around the 550 bet mark )it’s caused by 2 large losing back bets, a £1000 at Coral and a £570 at Tobet.
Overall in all the bets I’ve placed there are only about 50 bets that each generated a +/- £500+ figure which got me thinking.
If I was to ever try not laying my bets I’d definitely reduce my stakes in order to reduce the impact of variance (I appreciate this would impact some offers where I would normally do full balance bets but that is conversation for another time).
Let’s take a quick look at the profitability of back bets with a stake of over £100.
Isolating back bets with a stake of over £100 shows great profitability (nearly £10k from 320 bets) but also huge variance much as we’d expect.
It also looks like we may have hit a good run towards the end as the last part of the graph looks significantly steeper than the earlier phases.
If those last few bets had lost it would paint a very different picture.
We’ll now take a quick look at the bets with a smaller stake.
By reducing the stake we see a lovely profit graph, showing virtually continuous profit with very little fluctuation. Profit per bet is much lower than when the stake is larger but that’s to be expected with the reduced stake.
Analysis of my bets clearly shows that I would have been best off following an approach of reducing my stake size and not laying any of my bets at the exchange.
But what else could I have done?
High odds or Low odds
Something else I wanted to explore was to try and understand what impact the odds had on profitability. Again I’ve stuck with just using the back bets as it reduces the number of variables.
Here’s a brief summary of what I found:
What does this mean I hear you ask, well a few things stand out:
- Significantly less bets are placed at odds of less than 1,49 – this is to be expected as they normally don’t count towards wagering requirements
- Odds of 1.5 -1.99 seem to give significantly less profit per bet than higher odds bets. I’m not particularly surprised as I’ve always felt that bookies underpriced short price favourites.
- High odds (5.0+) seem to have extremely high profit margins, – I suspect this is due to the impact of SNR’s bets.
Removing SNR bets from the analysis seems to support this.
By removing SNR bets from the analysis we can see that with the exception of the 1.5 -1.99 range we get profit of 8-17% on average,
As Stake not returned bets and account adjustments make up the vast majority of my bonuses I’m amazed to see these profit levels when they’re excluded.
For a final odds comparison I’ll remove all the remaining bonuses (basically the stake returned bonuses) leaving just ‘normal bets’.
Before you see the results it’s worth highlighting that I aim to placed back bets at odds that I think are overpriced by using an odds comparison service to find back odds that are fairly high relative to the exchange odds.
I think it’s fair to say that the results were a surprise….
Can you guess which figure blew my mind?
There’s a lot to choose from but I’m going to focus in on the 10.29% overall profit.
Simplistically this means that after placing 2278 bets and staking over £150,000 I’ve averaged 10% profit from my back bets once bonuses are excluded.
What does this tell me?
For a long time I’ve known that I could profit using bonuses and laying off my bets.
I’ve suspected for a long time that by removing the exchange commission my profit would increase and I knew I was betting on out of line odds but I never thought the odds were enough out of line to generate pure profit without the aid of bonuses.
Could we tweak the approach to make these bets even more profitable without bonuses?
Short answer, is yes I think we could, my investigating different odds, sites and bet sizes I expect we could isolate a type of bet that performs better than all others.
I did consider delving into this in more detail but…
Tthis blog is about generating a mathematically proven advantage over the bookmaker, although these results appear to indicate that I’d have made significant profit without laying bets or being awarded any bonuses this is far from mathematical proof.
I may explore it again in the future but for now it’s straying a little close to pure gambling for my liking.
What’s next? – I need YOUR help
I hope that if you’ve read this far you’ve found this analysis to be useful. I’ve certainly found it an interesting exercise that I’d like to take further.
My data set is pretty minimal covering only 2500-7500 bets (depending on the analysis), I’d love to be able to replicate this over a larger data set and this is where I need your help.
I’d love to get a copy of yout betting records. I clearly don’t need or want username and passwords etc. I’m not even bothered about bookmaker names, what I’d like to see is the bets we’re placing as a community and the results of each of these bets so I can repeat this analysis on a larger data set.
If you’re interested in seeing this analysis redone on a larger scale or if you have any other theories you’d like me to test please drop me an email matt[at]theadvantagegambler.com.
How to send data to me?
The ideal for me is an emailed copy of your ‘betting log’ from the Ultimatcher spreadsheet. The more people who can send these to me the larger the dataset I’ll have to redo the analysis.
If you’ve reached this point, thanks for reading this far I know it’s a massive post but hopefully you’ve found something in it useful.
Let me know your thoughts in the comments below, particularly if there’s any other theories you’d like me to run through my data,