By Peter May
Systems are an integral part of betting. Whether it is horseracing, football, casino games or the National Lottery, if there’s money bet on the outcome of an event then someone, somewhere, will have a system which, it will no doubt be claimed, is guaranteed to make a profit.
In the last two decades the widespread proliferation of personal computers and the availability of sporting data have made it much easier for racing enthusiasts to develop their own betting systems. This has led to a better awareness of the relationships between the key racing factors and the outcome of races, putting the bettor in a much stronger position from which to formulate a winning strategy. Using horseracing data to inform decisions and design a well structured, systematic approach to betting remains the most reliable way to return a long-term profit from the sport.
Not every bettor would necessarily agree with this, and a great many firmly believe that systems are the poor relation of value betting and the more conventional approaches to solving the punters' problem. However these apparently diverse methods have much more in common than it would seem and it can be argued that everyone who bets on horses uses a systematic approach to a greater or lesser extent.
An analyst who studies the form and bases all decisions purely on information gleaned from historical races and statistical relationships may feel that a system would not form even the smallest part of the method. But consider this conventional approach to race analysis: the form is analysed, ratings checked, previous race times studied, suitability of the race conditions determined, strength of the race in terms of the opposition evaluated, along with running styles and the likely pace or lack of it.
These conclusions are combined into a price, a figure believed to represent the chance of success for each, or a selected number, of the runners. The final step in the process determines whether a bet will be made or not and essentially concerns the following test:
"If the price available is greater than the calculated value price then back the horse."
But isn’t that a system? An if..then.. rule, in this case featuring just one condition, is the definition a system. Whether the bet is made or not is determined by this rule. If the condition is met the bet is placed, otherwise it isn’t. Some may claim that the calculation of the value price is not systematic since it is based purely on judgement of the facts and the mere application of a rule at the very final stage of the process does not constitute a system.
This is a reasonable view. However the process of determining a value price concerns making evaluations based on the data relating to each horse. These assessments can only be made by the application of rules learned over many years of betting. Without the in-depth knowledge these connections provide it is simply not possible to draw any meaningful conclusions about a race. For instance, when discussing a bet a race analyst who adopts a non-systematic approach might say “I like it when jockey x rides for stable z”, or “I don’t like small horses around this big track”.
These are connections which have been distilled from historical race results. For the system user these rules will normally be the direct outcome of data analysis; for conventional bettors they are probably due to assimilation, by noticing the patterns as they occur and mentally recording them. Both rule identification approaches are valid providing they have been established in a logical, precise fashion. My preference, and that of the system player, is for the former. It is possible to be more confident about the validity of the rule if the data supporting it can be checked and analysed. But the outcome is the same for both learning techniques: a set of rules which forms a knowledge base that can be used in the analysis of future races.
For many bettors their reluctance to use systems is due in part to the perceived simplicity of such techniques. It is difficult to see how a rule with perhaps just two conditions could outperform a detailed analysis of a race involving all available information. However, it is important to remember that a system is not designed to predict the race winner, but instead is attempting to identify a small region of the market where the chance of success is not fully accounted for by the prices on offer. Therefore some rules can be simplistic in their form and still produce profitable returns.
For instance, many of my most profitable systems have had very few conditions, specifically those based on the draw bias. For certain tracks the rule would simply state that all horses running from a particular stall over a certain distance should be backed. The reason these methods worked so well had nothing to do with the accuracy with which they determined the chance of success. Their effectiveness was due to the fact that a critical factor, the stall allocation, had not been fully taken into consideration by those setting the prices, and therefore presented an ideal opportunity for making a profit.
The advantages of using systems over conventional race analysis methods are due to their consistency and unambiguous form. A system or rule can be applied just as well today as it can tomorrow. However, the analysis of a race can differ depending on the way the analyst views the event. This is due to the subjective nature of race analysis. A conclusion determined from a set of results today may not agree with one derived from an identical set of data some months ago. These changes may be a result of the performance of recent selections. It is easy for an analyst to change a method, perhaps unknowingly, if experiencing a particularly good or poor run. Over confidence, or lack of confidence, can play an important role in the analysis of a race and consequently remove any degree of consistency over an extended period of time.
For system players this is not such a problem since the outcome of a system is not subject to these influences and the determination of a qualifier will remain the same regardless of recent events. However a protracted losing run is likely to engender that same level of reserve. In these circumstances tried and tested systems become the objects of tinkering efforts to rectify the situation. The solution to this problem is rooted firmly in the system design process. If the method is founded in a logical manner, validated accurately on an independent data set and staked appropriately then, in general, there is little reason to question its likely outcome. Naturally there are cases where systems are no longer valid and should be omitted from a bettor’s portfolio, but these can be identified and avoided.
Essentially, given their form and structure, systems provide a consistent approach to betting and one which should form part of any bettor’s portfolio of techniques employed to make a profit. Though conventional race analysis can still provide a means of generating an income, running systems alongside this approach can significantly increase the size of the betting bank.
Methods for designing and validating systems, as well as other system related techniques are outlined in the following articles and like this section have been distilled from the book In Search Of The Winning System, published by Raceform, which charts my search for winning strategies over the past 40 plus years.