VALUE INVESTING
Which Form of Form?

By Stephen Bland (TMFPyad)
April 28, 2000

I'm going to look again this week at one of my favourite topics, odds, and why a failure to study them will increase your chances of losing money in shares.

In a football match there are three possible outcomes: win, lose or draw. If you knew nothing about the teams playing, and assumed they were evenly matched, you would have one chance in three of forecasting the outcome correctly. In bookie speak, the odds are 2-1 against getting it right. At 2-1 our bookie would break even over a large number of games. He would in consequence have to offer less than that to make money. That difference would be his margin. If he offered say 3-2, then he would win over time.

The same kind of thinking applies to casino games like blackjack, craps or roulette. The house always has to have an edge, has to offer somewhat worse odds than the true mathematical chance of winning the game so that in the long run they must win. For example, in roulette, the house advantage lies with the number zero where all bets lose. If you are betting on an even chance like black/red, high/low, odd/even, they will offer you evens, or 1-1. Bet £1 and if you win you end up with £2. Of the 36 numbers from 1 to 36 on the wheel there are an equal number of each of these even money bets. Thus 18 blacks and 18 reds and so on. If that were all there was to it, the house would come out even. No way to run a business.

So they have their zero, making 37 numbers in all. That slight advantage is all they need to make sure that bets offering even money, like black/red, produce long term profits for the casino. The actual figures for even money bets on roulette show that the chance of winning is 18/37. In odds parlance, this converts to (37/18 - 1) = 1.056-1. But in fact the win money they are offering you for getting it right is only evens, 1-1. The small difference is their margin. They cannot lose over time, although they can have bad days when they are down. In American casinos, it gets worse because they have two zeros, thus improving the odds even more in their favour.

Casino games are pure chance, there is no form to study. Each roulette wheel spin is entirely independent of all previous spins and you cannot infer any future performance from past results, just like they say in the fund ads. But what of the football match I mention above?

Suppose that instead of two teams unknown to you, which I presumed were of equal ability, we have a match of unequals. Looking at London, say a cup tie between Arsenal at home and Leyton Orient. A major mismatch, on paper. What are the odds now? Well, we have a top premier league side against a third division team. On a pure mathematical approach any of the three outcomes remains a 2-1 shot. But the actual odds offered by the bookies would be wildly different. Arsenal to win would be on a very low, likely odds on, situation. Orient to win or draw might be quite long, way over the mathematical chance. So does it follow that Orient to win at say 10-1 is a great bet, because in theory they have a 2-1 chance of making it? No.

Because here of course is where form comes into it. Unlike roulette, where form cannot exist, punters will believe, rightly, that the team far higher up the football ladder stands a massively increased chance of winning such a game for historical reasons of how these teams have performed in the past, their standing in the league. Therefore, they will put their money where their beliefs lie and consequently most of the stake will go on Arsenal to win, only a little on Orient to win or draw. The odds on Arsenal to win will be driven down by the weight of the money bet on that outcome. But the trouble with such a form-derived bet is that you cannot calculate the odds yourself. You can calculate the simple mathematical chance of 2-1, but you cannot put absolute figures on the real odds based on form so as to know whether it is worth risking money or not. But you can have a go: merely betting on Arsenal rather than Orient is more likely to produce profits.

Okay, I'm saying that some gambles have a clearly defined mathematical picture, like roulette, where past results tell you nothing of the future. Others, like a seriously mismatched football game, are unclear in what the exact situation may be, because past results have a lot of effect, but you cannot quantify this easily although you have to try. Similar situations to the latter arise with horse racing where the actual odds for each horse will be shorter or longer than the strict mathematical odds in which form is ignored, because of each horse's differing form.

And of what of shares? They are closer to the football and racing situation than they are to roulette. You have to study form if you want to shorten the odds in your favour. Form in this case being whatever techniques you have developed which seem to work, that is fundamentals, technical analysis, DCF/NPV and so on.

But people don't. They regularly take mad gambles, like betting on Orient to beat Arsenal, without any knowledge of form whatsoever, or sometimes knowledge that is unreliable, form of a sort, but not the form that produces results. Vague stuff like "people have to eat, therefore supermarkets must be good shares". Somerfield (LSE: SOF) shareholders would disagree violently. Or "everybody has been buying some share, therefore it must be good", even though the investor may not know the first thing about it. We see this kind of poor, unfocused reasoning all the time. Homing in on some pet belief rather than the underlying qualities of the share itself. Even if the pet belief is true, people having to eat undoubtedly falling into that category, it doesn't mean that a particular share will do well out of it. Occasionally it will; occasionally Orient would beat Arsenal, but the absolutely fundamental issue is that this will never happen often enough that money can be made regularly over long periods.

And what has this to do with value shares? Just this. To win you must study form, in my view. But more than that, you must study the particular form of form which has been proven to have a high success rate. There is no point in devoting yourself to an approach requiring such study unless you are pretty convinced that it will pay. My view is that value is an approach that works very well. Out of all the many techniques that people use, this has to have among the highest repeat success rates over the long term.

I believe, through experience rather than scientific proof I admit, that the odds of winning with value are turned heavily in favour of the investor, in contrast to many other styles. The approach is offering me win money which is better than the chance of winning. Beating the bookies if you like. Equally, and a crucial part of the strategy, the punishment for getting it wrong is likely to be pretty mild because of the extensive safety net features. I say crucial because with many other investing styles substantial losses can occur for getting it wrong. This will make a big hole in long-term performance. With very deep value, there are still losses sometimes, but they are likely to be much softer than say a growth or NPV investor who gets it wrong. Reducing losses is one of the first steps to successful investment.

But all this comes at a cost. Research, of the right kind, is essential. Patience in finding suitable shares, then more patience in hanging on as long as you have reason to believe in them, is essential. Being willing to be out of the market because you can't find qualifying shares and not weakening your criteria to go into something inferior is essential. Selling is essential as soon as you feel the shares have lost a lot of whatever it is took you into them in the first place. Always better to sell too soon, always better to leave something for the next guy. The higher risk bit. I seen it mentioned that selling too soon is a common fault of the little guy. One of those vague generalisations again. With my value shares I'd say the opposite is true. I nearly always sell too soon. I love my cash too much.

Related Links

• Value Investing discussion board