Recently in Major League Baseball in America, the Oakland Athletics were on a roll. They had won 25 of their last 30 matches and had penned themselves in to receive a good position in the playoffs, in little over a month.

However, few could have predicted what happened over the following couple of weeks. They went from being almost unbeatable to be the struggling side of yesteryear, losing seven of their last eight games. Of course there is major publicity. The media are asking what has gone wrong. Maybe it's the coaching? Have they become too cocky? Many theories were floating around.

But statistically, this is not out of the ordinary. Whilst it might be that Oakland had indeed improved, many statistical papers have proven that the big run and consequent losses were simply the result of random variation.

Gamblers rarely understand the idea of random variation. I remember when a lecturer at university gave me a great example. He gave us two pieces of paper with thousands of dots printed on each of them. The thing was, that on only one of them the dots were randomly placed. When looking at the sheets the one that was not random looked the most random! Dots seemed to be evenly placed everywhere, whereas the one that was random didn't look random at all. You could see patterns, clumps of dots and wide open spaces where no dots were printed.

This is the effect of random variation. It won't always go to plan. The example of Oakland's winning streak can easily be applied to gambling. If we were to win 25 of 30 winning bets and then lose 7 of the next 8, would we decide to stop gambling because of the downturn? We might wish we had stopped when the going was good because a downturn was shortly and undoubtedly going to occur.
This unfortunately is often the downfall, not of the betting results, but rather the punter. Many punters decide to stop betting when the going is rough, or amazingly decide to stop betting when it is good because there will no doubt be a downturn around the corner. By the same logic, does this mean that one should start betting when we are doing badly because an obvious 'upturn' will be around the corner?

Unfortunately no . . .

All this is merely a result of random variation. Of course there can be outside influences. The tipster is not as good as previously, the game has changed and there are other variables not taken into account, but most of the time it is purely due to random variation.

We saw in the Punting Ace's NRL 2005 model exactly that problem when you jump on and off the bandwagon. At the start of the year, the NRL model was on fire, posting big weekly profits in the opening five rounds. Shortly after however, we had a losing streak and some people decided to stop betting. Those who stuck with it were rewarded with an overall profit for the year.

In fact I was so intrigued about the problems of bandwagon jumping, I decided to do a simulation to see how much of a detriment it can be to your gambling.

I set up a simulation where we make 10,000 bets, all at odds of $1.95. Using a $1,000 bank and betting $50 each bet our success rate is 53%. This means that we should make a profit of almost $18,000 or a return on investment of 3.6%. And we did that when we were betting on every single bet.

However for the bandwagon jumper this is different. I set it up so that if the gambler had lost $500 or more in the last 30 bets then he would stop betting until he can see that the gambling model was doing well again. By doing well, I decided that he would jump back on when a profit of $1,000 was made. In other words, it would make back the $500 and show another profit of $500, ensuring to the bandwagon jumper that it was a good time to jump on. I set the simulation to bet 5% of the bank ($50 of $1,000).

The results were amazing. The bandwagon jumper, jumped off 13 times over the 10,000 bets of the simulation. This means that even though we were betting 5% of our bank, there were stretches, many of them, where we lost $500. The bandwagon jumper jumped back on all occasions, but had missed some good betting in the meantime. In fact when looking at the results, the bandwagon jumper recorded the same percentage win rate and the same return on investment rate as the constant gambler, however because of the drop in turnover, he didn't make as much. In fact, the bandwagon jumper made only 40% of the bets that the constant better made, which essentially meant that he made only 40% of the profits as the gambler who gambled on every single bet.

Bandwagon jumping is not good for your gambling, but some would argue that it is needed for success when dealing with a method or tipster that is not very good. My response is that you really shouldn't be on a method or tipster in the first place that cannot prove his long-term success. A lengthy and confirmed betting history is the only way to prove the long-term success of any gambling method or tipster. Once you've proved its long-term success, don't jump off and on, stick to it through thick and thin, and you will reap the long-term rewards that every consistent professional gambler receives.

This article is protected by international Copyright © Elk Publications Pty Ltd October 2004 Please contact if you wish to reproduce this article elsewhere.