This post originally appeared at www.daily25.com and has been reposted here with the permission of Steve from the Daily25 blog. It is written by Matthew Trenhaile, who has worked as an odds compiler for many years and is now on his own taking on the bookies. You can follow Matthew on twitter @CrazedAlchemist and he also has his own blog. Over to Matthew.
Matthew’s Background as an Odds Compiler
My name is Matthew Trenhaile, and I have worked for six years at the UK Sport Spread Betting division of the IG Index, as an odds compiler. Steve was kind enough to allow me to utilise his blog to show my abilities in writing, through a series of articles from someone who has actually worked with a major bookmaker and in the industry. To begin with, I’d like to address some phrases I am going to use continuously throughout these articles, as the names and nicknames for different situations are different from nation to nation;
Jolly = Favourite to win
Rag = Underdog to win
Sharp = Smart punters
Square = Losing punters
Books = Bookmakers
Bowler = An account in another person’s name, used to place bets
When I describe my odds it will be in decimal (European) odds, and when I am referring to an amount of money, the currency will be Pounds (UK).
A lot of the articles I have read lately, in addition to television pieces, have generated a lot of debate. This is completely fine by me, as long as it has good intentions. These are my opinions, and I respect the ones of others as well.
An Evolving Betting Industry
In all of the following and future articles, I’ll be attempting to tackle subjects from a past, present and future perspective. To begin, I will take a look at odds compiling and how this has changed over the years, with my main focus on the last fifteen years. I will also be looking into how bookies copy their prices off each other, which is a topic that is relevant in the industry today. The industry concerned with Spread Betting stands responsible for every change that is of any significance in odds creations of the last 25 years. This is also where my perspective and experience comes from.
I would advise readers that aren’t familiar with Spread Betting to read at Sporting Index’s site and go through their training section to get an idea of what’s involved in this. In sports Spread Betting you could oppose an outcome before you could lay on Betfair, and you could bet in-running online at the spreads before any of the fixed odds bookies. The doomed to be a failure product Extrabet, with the dreaded close out button, was also created by IG Index. Even though this doesn’t necessarily mean that I am extremely great at compiling odds in comparison with others, it definitely means that I was at the centre of advancements in technology in the sport betting industry. This was the reality back then, and the sports spread betting industry is still the leader. This goes through the employees, who devise the models of other bookies. It can also happen through firms such as Sporting Solutions, a spin-off from Sporting Index, where in-running prices towards bookies is the product. It is a clever move by fixed odds bettors to take a look at the spread betting firms prices, as their secondary check or backup before actually placing their bets.
How Odds Compiling Works
Databases, Statistics and mathematical models all play an increasingly large role in compiling odds, rather than personal experience, intuition and feel. When I worked, I was fortunate enough to work alongside odds compilers that worked in the “old” way, and had watched thousands of hours of horse racing, and could distinguish the smallest of nuances of how the horse was ridden and the strategies of the trainers. I also worked with people who broke down sports into their fundamental inputs and turned those inputs into probabilities. This was done before and ruing races.
Initially, odds compilers were split on the prospects of Betfair, and especially on its uses with regards to compiling prices. When I started working, Betfair had only just reached liquidity levels that were significant, and couldn’t be ignored due to the fact of the increasingly large number of arbitrage bettors. After 6 years, we priced all horse racing products after our Betfair API, only using one person to oversee that the process was correct. This was an extreme contrast to back when I started, where we had one trader for every horse race and a room filled with 40 traders, even though the number of sporting events and matches were only a tenth of what we have now. As a result of being a subsidiary of a large financial firm, we were paid more than the remainder of the industry, and our resources topped every other bookmaker.
Our resources were ploughed into trading at more events in-running and to develop more complex models to generate odds in-running, at an increasingly high rate. Generating these odds provided us with liquidity to the betting exchanges, and in turn generated a significant secondary revenue stream for us, in addition to providing these odds for our clients` sake.
Creating Statistical Models for Sports
The statistical odds compiling mostly consisted and originated from the counting of how often an event had happened previously. If we were compiling data from two football teams, we would look at how many times the home team had won in their last 20 home games, and how many times the away team had won in their last 20 away-games.
This was prior to my time. In the old days, the edge was found by bookmakers, simply by studying the game more than the bettors and comparative odds knowledge. In other words, if a bookmaker found a 50% chance for a team to win its home games, and decided to put up a 1.85 odds, the bookie could trick the punter the next time, by putting up 1.75 odds. This is just because the bettor remembers what they played, and places the same bet the next time, regardless of what the odds are. The bookie will then have extracted value from the bettor, just by knowing what punters have played recently.
Odds compiling used to be more focused on the bets of the punters, rather than the probability of the outcomes. To this date, this is the very difference between bookmaking and punting. The bookies` job is to understand where money will flow, as opposed to the punter, who has to understand the probability of outcomes and recognise value.
In the world we live in today, the odds will reflect probabilities of an outcome more and more precisely, and is lesser concerned about the opinion of the public. The upspring of in-running betting is what stands behind odds compilation through mathematical modelling. This meant that it became too difficult for people and compilers to recognise prices in multiple markets for multiple events in-running, only through the use of a pen and some paper, rather than computers. Bookies had a need for automatisation, preferably through models.
The Possion Distribution for Modelling Sports
Almost every model for sports betting can be found on the internet, and have been available for quite some time. These models have been improved over a number of years, making the data greater. “Poisson” distribution led the way as the best football-model, because of its accuracy, as a result to being improved, in addition to being easier to add time decay to the inputs.
Each team has their goal inputs, and for every minute the game goes on, the model decreases. This means that as the simulation continues, the odds is recalculated, and changes a little bit every second, even when not significant enough to be visible. In the beginning, Poisson distribution was the only way, but we later changed to custom distribution for every league, and from goals scored to the expected number of goals based on shots.
These days, compilers measure the overall quality of the shots being taken, in addition to measure the individual players impact on the shots, to create a better understanding, and to create highly improved shots-models. Or are they actually doing this?
Every sport has a model that’s similar to this model, and all can be improved as the level, amount of information and data are made obtainable. The important point to focus on for bookies, is whether or not they wish to go down this road. Do they go through with the payments to secure this information, and do they pay people to work for them and maintain it? Or do they hire someone else to do it for them?
Multiple firms are already in use of the same price from the same provider of in-running football, for example. How good would your odds need to be to beat the average punter if you only used in-house compiling? Can great management of risk hide a multitude of sins just by letting the best of the punters move the prices so that they have the best outcome for you?
It is sad to say, but only a small number of bookies allow that style of management of risk, or for the investment in greater pricing. Compilers are now understanding that a pound spent on advertisement and marketing can initiate a greater profit than the pound they could have invested in improving the quality of the software or on staff.
In general, any bookmaker has a goal of beating the 98% of all punters with the lowest number of staff possible. If we go into greater detail, bookies are most cost efficient when hiring young and inexperienced staff, in addition to give optimisation of software little or no thought. This is seen as a better option than hiring experienced odds compilers.
How Bookmakers` Business Models and Views on In-House Odds Compiling Has Changed
You might be convinced that since odds are the product served by bookies, they would be better off with focusing on developing their product. However, in the world we live in today, that is not correct.
For example, you wouldn’t base the choice of your hotel based on the cost of a beer in the bar compared to other hotels. Bookies look at odds at about the same level of importance as the price of that beer. The real product they are selling is entertainment, and not the intellectual contest between the bookie and the punter. It would be foolish to think that they have ever sold the product of this sort of competition. The product bookmakers sell is a rush of adrenaline, and those who still believe that characters competing with punters, and playing against each other in betting rings concerned with horse racing is what betting used to be, are greatly mistaken.
All punters have a different story and feeling towards how sports betting and bookies used to be. People I have spoken with remember the taxes in the UK, large margins, only a choice between a few bookies, refusal of payment when trying to withdraw money, and in some cases threats of violence when trying to get paid. I do not know where the image of a gentleman-bookmaker comes from. Over the years, bookmakers have taken bets and they have refused bets, they have allowed high stakes and they have refused them, and some have filed for bankruptcy and some haven’t. This has been the truth for square and sharp punters.
Every bookie expects to win in almost all cases, no matter what the quality of their odds are, and to improve channeling of their efforts in getting punters in the door, rather than having a cheap price in the bar when buying beer. I do not care much for this model, but I believe it is going to stick.
There Can Only Be One Pinnacle
Now, let’s dive into the compulsory paragraph about Pinnacle. Why doesn’t every bookie operate such as Pinnacle?
Well, the simple answer to this question is that there can only be one company at a time with the business model that Pinnacle possesses. Some might use parts of it in some areas of their business, but the model that Pinnacle uses isn’t easily duplicated.
Pinnacle’s tagline is that they openly welcome winners, and has stated that this is due to their wish to beat all square book with the best proprietary trading models that the market can offer.
Having the greatest employees (including odds compilers), betting on tiny margins, welcoming arbitrage- and sharp traders, in addition to changing the odds according to how sharp they are, the odds at Pinnacle will be the best representation of the probability of any outcome at sporting events.
Start with your limits low, and increase them according to the increase in your price. Not a penny will be spent on marketing, but everyone’s allowed to use the affiliate banners, in addition to giving your API to everyone you can. This includes arbitrage and services that compare odds. The low margins result in you being at the other side of arbitrage trades. The square bookie your price is up against is so bad at placing the price that you must be getting the value side of the arbitrage bet.
Actually, you will be getting this scenario again and again, but will always end up moving your price to make sure that you get every bit of value, every single time. This ends up with you becoming the greatest punter that William Hill, Ladbrokes, Bet365 and every other bookie has seen, without the need of opening even just one account. The risk management group that are in your possession are almost fully automated, and you need the services of an investment bank. To be able to set up and use this kind of model, they must be doing so with smaller margins than you, in addition to having a greater risk management and a lot of capital. The only example of someone giving this a shot at the moment, is Marathon, and they are still nowhere near the point where they can afford to keep winning accounts in their books.
But what about the Asian bookies that keep accepting winning punters?
Well, as long as they see a great amount of money coming in, they are happily accepting them to play. This is the case in football, certain American sports, and in the more known tennis match-ups.
The Asian Handicap Model was designed to function at a low margin, move quickly, have a great volume of trades, in addition to not working with the sharp money. In some cases, it was designed to play at early and low limits, just like Pinnacle, while the market still forms. This should result in the risk being low.
Still, they don’t directly run a Pinnacle Model, but they have large enough square volume to use some of the same features of it, and handle a loss comfortably. This is IF the books as a whole, result in a positive profit.
The Future of Odds Compiling
In the end, I think that compiling of odds will strictly become an outsourced function for bookies. Companies that specialise in compiling odds, both pre-match and in-running, will arise. Already, multiple companies provide this option, in addition to there being companies that provide liquidity to betting exchanges. Homogenised pricing is also something that will occur more often.
The reasoning for why you see a lot of similar pricing in today’s market, is that the liquidity in certain sports betting markets is of such a high volume, that we can see something that reminds us of a market price, similar to the financial markets. It doesn’t make sense for bookmakers to have deviating prices from each other in football that can be categorised as top-level. Trust me when I say that substantial studies have been completed, and that they conclude with the fact that trading against the market in a liquid sport is a loss in the making as a bookie. When you are going to provide thousands of odds over a very high number of events, it is easier and better, to simply just follow the market.
Though, this might not be the case for a punter.
An increasing amount of companies will look at a market and tell you that the combined price from a liquid Betfair, Bet365, SBO, IBC and Pinnacle easily can be put together, and function more than well enough. These operations are not going anywhere, and see a very high amount of trades, which helps maintain the prices at a correct level.
“Why do they put up arbitrage opportunities, even in the liquid markets?”
The answer is that we are in a period of time, where the industry as a whole is in transition and sees changes. Some are still trying to understand and accept that the price of the market should be set from the flow of money, rather than the opinion of a trader.
I predict that bookies will have a growing confidence towards these consolidated feeds. They will eliminate arbitrage opportunities, or at the very least, with sharp sources, they will be cheap in use. I also believe that bookmakers will outsource all of the risk management tasks. There will forever be exotic markets attached to the core liquid ones, and if the bookies have any sense they will increase the limits at the core ones and decrease the limits at the exotic. This should be at a level of a 100 pounds takeout, which is a great amount for punters that are in this for the entertainment of it.
When focusing on horse racing, I have a few predictions. It might go in the direction of Starting Price only. I actually think that pari-mutuel betting with very low margins could be the way forward, making it into a universal Betfair with Starting Price. This type of pool betting has not been prevented in Asia, not with punters or professional bettors.
The problem that we have with the UK is the high amount of bad racing. Who knows what would have happened if bookies had pooled tickets, and if Betfair and the Tote had shared the profits, maybe it could have resulted in something valuable for everyone. This could potentially put to bed some of the bad Starting Price rigging that is happening.
Betting will become just as much about beating the market as a whole, and not picking off the sick and the weak prices from the books. This will lead to a requirement towards punters and tipsters to understand the weaknesses and the strengths of the markets they work in, just as much as the sports themselves.
As a final word, I would like to say that you shouldn’t concern yourself with how the prices and the odds are set and compiled, or whether bookies show the same prices, as the only thing you should concern yourself with is to beat them. There will always be flaws and edges to be found in every sport, even though they get harder to spot, and shrink in size.
Written by: Matthew Trenhaile
Follow Matthew on twitter: @CrazedAlchemist
Read his blog here.