If you don’t want to listen, you have to pay?
So, in 1990, I had the ultimate idea of how to approach the problem of predicting the probabilities of football matches in principle. The programme had also matured quite a bit in all-night (because I still had my SEL job during the day) work. The algorithms had been implemented so far, the calculations worked, the essential functions were ready to run. But the quality of my programme and its predictions, despite the initial “success” at the 1990 World Cup, still had to be proven in the long term and in everyday league life.
So what was to be done? My friend Abi Rosenthal, who had long been a millionaire and had also helped me in Monte Carlo (or was it the other way round?) believed in me and … simply founded a company. He had experience in that. It was a limited company. There had to be 50000 DM capital, we were equal partners and I suddenly had 25000 DM capital, my contribution. Of course, Abi had also put in this part. But I owed him the money, at least in theory. If the company failed … but we hadn’t dealt with that.
He rented me an office in the centre of Berlin, a new, better computer was bought and I was allowed to do what I wanted, so to speak. But I worked, seriously worked. The programme was improved little by little. Statistics were compiled, studies of past seasons, theoretical forecasts, etc. were made. So I already had estimates for a lot of European leagues. And what was I supposed to do with them?
Personally, I ran the whole system automatically at that time. When I decided to quit my really good job at the SEL, I didn’t know definitively in which direction I would go. I had my programme, it was good but not yet tested (enough). And I had the first good results. I could have imagined anything: So providing bookmakers with odds, selling my programme or even becoming a bookmaker myself, opening a betting shop. That was actually one of the main reasons for quitting.
The latter has become more and more superfluous, as the state has permanently and consistently put obstacles in the way of betting providers, especially in Germany, until today (2009). The reasoning is clear to me: preservation of the betting monopoly. Although I am of the opinion that this is an own goal, since gambling can be done worldwide on the internet and the state thus only loses the tax revenue it could earn by gambling, the reason for this is actually only ignorance and a lack of understanding of the correlations. I also wrote this book to clarify this.
So I started to record all the odds of 8 bookmakers week by week. I wrote a small programme for this. Then the computer compared the odds with another function: My computer generated odds with the odds offered by all the bookmakers entered. Then, bookmaker by bookmaker, it produced a printout where all the bets that were favourable according to the computer, i.e. my findings, were recorded. On it was the fair odds, i.e. the correct payout odds in my estimation as the reciprocal of the probability and the odds actually paid by the bookmaker. So every game that the computer advised me in this way was theoretically bet. For the sake of simplicity, I bet each game monotonously with one theoretical unit. I sent the whole thing by fax to these 8 bookmakers before each weekend, i.e. before the end of the games.
After the end of the weekend, another function evaluated all these theoretical bets. And there was a result for each of these bookmakers. I then faxed this result, favourable or unfavourable, to these 8 bookmakers again on Monday.
I did this to check myself, but also to prove to these providers that my system would be profitable in the long run. Of course, I had no way of knowing yet if it would work. But it seemed to work, the results were good.
I still sent my results to each bookmaker after a few weeks had passed. That way they could get an idea of how I was doing overall. For me, it was the means to check the overall effectiveness of the system. And I had the hope that the bookmakers had to assume that these results were not falsified, since they could at least check my results with them themselves.
After half a year, I had a result that gave me enough information: I had won with all bookmakers. At one more, at the other less, but everywhere I had made a profit. This profit was about 5%. The good thing about it was that it was all achieved with single bets (for optimisation with system bets, please study the chapter “My system”). What’s more, with one unit per bet. So there was neither the exploitation of the advantage of increasing the advantage with combination (or in my case system) bets. Moreover, I have not yet used the effect that I could even play particularly good games more expensively, i.e. higher.
Even a third effect could have a positive impact: Namely, that the computer had not yet filtered the odds. For example, it got odds of 4.50 at one bookmaker and played that, provided the fair odds were below that. But if there was even a 5.0 elsewhere, it would have been sufficient to play the bet only for the 5.00, that is, for the best rate. That would take into account the eventuality that the fair rate was actually above 4.50 but below 5.00. Because I’m definitely not omniscient.
However, the results were so satisfactory that I had definitely expected positive reactions. But I had overestimated the effect of my faxes. Maybe they weren’t even read or looked at? In any case, I made a few phone calls without any particular results. SSP would have invited me to England and received me there, yes. Sportwetten Salzburg had expressed interest and even suggested a meeting. But that was much less than hoped/expected. One still had to feel in the beggar’s position. And I really don’t like that. But who would? I also didn’t feel called to be a hoover salesman or do any other “door-to-doors”.
I now had enough confirmation that my programme and the algorithms were good. So I asked Abi to lend me one of his cars, packed my computer and the other six things and set off, with Salzburg as my first destination. After that, I wanted to visit a few other betting shops, in Vienna, Innsbruck and Linz. In Austria, betting was not only allowed but even popular.
The travel report can be found in the chapter “Sports Betting Salzburg”. Here is just this much to add: The only betting office interested in my odds was the Vierklee betting office in Innsbruck. From then on I supplied them weekly with the computer odds and received about DM 400 per week.
However, the amounts earned from this were obviously not enough to finance my living. So the decision was even easier: I could get the 5% in cash instead of virtually. And with the possibility of combining games and thus increasing the advantage, that was merely the lower limit. Let’s get to work. I asked Abi for a betting budget. He agreed immediately, opened a new account for me with an initial balance of DM 25,000.
And practice confirmed the theory. After about two years, Abi got back DM 100,000. And that was only his share.