How to assign Stakes to a Tipsters portfolio

Next we are going to expose, as we had promised in another previous article, a simple model to assign stakes to a portfolio of tipsters. The objective is not, far from it, to defend this model as unique and irrefutable, but to help the beginners in some way to manage their bankroll, so that they do not fall into certain errors that are quite common.

Suppose I have a bankroll of € 10,000 and the following tipsters, which, after an individual analysis, I think are reliable and I would like to follow:

BetsFutAmateur: 19% yield. Drawdown: -34 Uts . Stakes: 80 Uts / month

PickNick: 15% yield. Drawdown: - 18 Uts. Stakes: 150 Uts / month

TheManInTheKnow: 20% yield. Drawdown: -120 Uts. Stakes: 160 Uts / month

* Remember that the drawdown is the maximum historical drop of the tipster and serves as an approximate way to measure the variance or irregularity in his results.

Many people decide, as a standard strategy, to place stake 1 equal to 1% of the bank, whatever the tipster. This has several negative consequences:

• If a tipster like PickNick (following our example) place many units per month, either because he has a very high average stake or because he sends more picks a month than others, you will be overexposed to this tipster and, perhaps, he is not the best tipster for you.
   
• Or, for example, if you have a high odds tipster like TheManInTheKnow, with a high drawdown, you could lose the whole bank in a bad run. The drawdown of TheManInTheKnow is -120 units. Losing 120% of the bank is going to bankrupt.

Under the higher statistical hypothesis that "the past is the best predictor of the future" and that, therefore, the yield, number of units wagered and drawdown will be similar in the future, we believe that we must take them into account when allocating the stakes.

There are many ways to approach this issue and there are people who do not care about variance or risk and what they want is to maximize profitability.

In our case, our objective will be to allocate the stakes in such a way that we maximize profitability by limiting the variance in the bank and trying to make it as small as possible.

For this, one option is to set an acceptable loss limit for tipster in your bank. It could be 30%. It is true that if bad runs happen at the same time, you could see your bank in trouble but the probability of that happening may be of the order of 0.0001%. One thing is that the three tipsters have slight losses at the same time and another thing that the largest historical loss of the three separately coincides in time. From this level of maximum loss we can already calculate the maximum value of the unit that we can allow.

• In the case of BetsFutAmateur, if his drawdown is -34 uts and we can allow as much -30%, your unit as much should be 30% / 34 = 0.89% of the bank. This is an approximation. In order to avoid complications, we would round up and assume in this case a unit at 1% of the bank, so € 100. If the drawdown occurs, we would be losing 1% * 34 units = 34% of the bank. Something above what we wanted to assume initially. With this unit, we would be talking about playing € 8000 per month
   
• In the case of Picknick if the drawdown is -18 uts, similarly it can be deduced that the maximum to place would be € 150 / unit. This would mean betting a monthly volume of € 22500
   
• In the case of TheManInTheKnow, like any tipster of high odds, things change. You have to be careful since his drawdown is usually very big. Doing the calculation we can get € 25 will be the maximim unit. In this case it would be € 4000 placed per month.

So far we have calculated the maximum value per unit that we could assign to each tipster so that if his drawdown is repeated, we should not lose more than 30% of the bank with him. But, what happens if we place this stakes?  The monthly amounts placed to each tipster are very unbalanced. Betsfut 8k, picknick 22.5k and themanintheknow 4k. This means that we are very very exposed to what picknick does. If he does it well our portfolio will do very well and if not, even the others do it well, their results will not have much weight in the global portfolio.

Let's give an example:

If picknick makes a slightly negative month with a yield of -1%, that implies -1% * 22.5k = -2.5k of loss. If the other two tipsters end the month in positive yield of 10%, we will win € 800 with betsfut and € 400 with themanintheknow. Therefore, our portfolio profit will be - € 1300 despite having two good and one slightly bad tipsters.

Therefore, here we must decide, depending on the confidence we have in each one, whether to assume these stakes or make some adjustment to balance weights depending on the confidence that we have in each one. It is a matter of confidence in the tipsters and the aversion to risk that each player has. There is no science. If I have blind confidence in all of them and I prefer to maximize the yield despite of variance, the ideal is to play those amounts and if picknick comes out a bad month, endure the loss. If, on the other hand, I have the same confidence in the all, but I do not want to be exposed to a picknick results, maybe I will try to balance the monthly amounts a little more. For example, dividing the picknick unit by 2, that is to say € 75, we would be placing € 11250 / month, with which we are getting closer to the objective of weighing them similarly.

In summary, it is necessary to consider when determining stakes the units played per month, the yield and variance, measured through the drawdown. A good method is defining the maximum loss allowed by tipster and then adjusting the monthly amounts so that there is no overexposure to a tipster. However, this is just one more method. Any other method that provides rigor and limits the most common errors previously explained, is equally valid.