From time to time, bets will be lost. It is inevitable since no selection system is 100% effective.

What happens next sorts out the professionals from the amateurs and the successful from the unsuccessful.

The professionals and successful punters recover their losses in the normal course of events. They use their staking plans and selection systems to recover their losses through their normal betting activity.

They do not change the way in which they operate their systems.

They do not invoke special procedures following a loss.

They simply go on about their business as normal and view their losses as part and parcel of their everyday life.

They view losses as an inevitable fact of life.

In my opinion, if a system is not capable of recovering its losses in the normal course of events, then it is better that the system is dispensed with.

Any selection system which requires a loss-recovery process to be invoked whenever a loss is encountered isn’t worth its salt in my opinion and should be dispensed with.

If a loss-recovery process needs to be invoked in order to maintain the profitability of a system, then it implies that the system, in its own right, isn’t profitable.

Such systems beg the question, is it the selection system or the loss-recovery system which is responsible for generating the profits?

At this point, some would state, ‘what does it matter if a profit is being made?’.

Well, it does matter and we shall now see why.

Loss recovery involves the placing of a larger than normal bet following a lost bet.

The stake involved is such that, if the bet wins, the losses are recovered.

The stakes are then returned to their usual level.

There is a variation of this method. It involves recovering the losses over several bets. Three is commonly used.

This method has the advantage that the stakes are not required to be increased to the same extent as they need to be if the losses are to be recovered through a single bet.

However, the variation has the disadvantage that all three recovery bets need to be successful in order that the previous losses are completely recovered.

Loss recovery processes rely on the premise that lightning will not strike twice in succession.

This implies that a loss-recovery process has a unique, even mystical, property.

It is as though there is something ‘special’ about a loss-recovery bet. It is as though recovery bets cannot lose.

Well, trust me, there isn’t anything special or unique about a loss recovery bet. It is subject to the same laws that normal bets are subject to. As such, they are no more likely to win or lose than normal bets.

Ah, there is one thing that I forgot.

They are different from normal bets in that, because of the vastly increased stakes, when they lose, they can prove to be very costly indeed, unlike normal bets. If they lose, losses can quickly escalate.

Most loss recovery systems are based on the Martingale principal.

This principle involves maintaining a consistent stake until such time as a bet is lost.

At this point, the stake on the next bet is increased in order to recover the losses incurred on the previous bet.

If this bet loses also, then the stake is increased yet again in order to recover the losses incurred on the previous two bets.

This process continues until such time as a bet is won. At this point, all of the funds lost are recovered.

This sounds as though it is a perfectly good and reasonable strategy to follow.

In fact, on the face of it, the strategy must surely win since, eventually, a winning bet will be encountered.

Well, I have had personal experience of this. It most certainly can fail.

In order to illustrate just how dangerous this strategy is, let us look at the following example:

Firstly, let us consider the system from a layer’s point of view: Suppose that the aim is to obtain an average profit of £2 on each bet.

For the moment, the betting exchange’s commission on all winning bets will be ignored.

Suppose that the first selection of the day is layed to lose at odds of 5.0 (4/1).

The stake, in order to win £2, must be set to £2.

The liability on the bet, should the selection win, is, therefore, £2 x (5 - 1) = £2 x 4 = £8.

If the selection loses and we win the bet, we have achieved our objective of winning £2 per bet and we simply move on to the next bet.

The stake on the next bet will be the usual £2.

If, on the other hand, the selection wins and the bet is lost, we will have failed in our objective to win £2.

In addition, we will have lost £8. We will have therefore failed to meet our objective by the sum of £8 + £2 = £10.

Now, let us consider our second bet, given that our previous bet lost: Suppose that the second selection of the day is layed to lose at odds of 5.0 (4/1).

On the first bet, £8 was lost. We also failed to make the target profit of £2. We also need to make a £2 profit on the second selection.

Therefore, the stake on the second bet becomes £8 + £2 + £2 = £12.

This is six times the normal bet!

If the selection wins, and we lose our bet, we will lose £12 x (5 - 1) = £12 x 4 = £48.

If the second selection loses and we win our bet, we will win £12.

We will have recovered our losses of £8 on the first bet and will have achieved our objective of winning £2 per bet.

We will therefore simply move on to the next bet.

The stake on the next bet will be the usual £2.

If, on the other hand, the second selection wins and the bet is lost, we will have failed in our objective to win £2.

In addition, we will have lost a further £48.

We will have therefore failed to meet our objective by the sum of £48 + £8 + £2 + £2 = £58.

So, in an attempt to recover the initial loss of £8 and achieve our target of winning £2 per bet, we have lost a further £48.

Now, let us consider the third bet, given that the previous two bets both lost:

Suppose that the third selection of the day is layed to lose at odds of 5.0 (4/1).

On the first bet, £8 was lost. On the second bet, £48 was lost.

We also failed to make our target profit of £2 each on the first and second bets.

We also need to make a £2 profit on the third bet.

Therefore, the stake on the third bet becomes £8 + £48 + £2 + £2 + £2 = £62. This is thirty-one times the normal bet!

If the selectionwins, and we lose our bet, we will lose £62 x (5 - 1) = £62 x 4 = £248.

If the third selection loses and we win our bet, we will win £62.

We will have recovered our losses of £8 on the first bet, £48 on the second bet and will have achieved the objective of winning £2 per bet.

We will therefore simply move on to the next bet. The stake on the next bet will be the usual £2.

If, on the other hand, the third selection wins and the bet is lost, we will have failed in our objective to win £2.

In addition, we will have lost a further £248. We will have therefore failed to meet our objective by the sum of £248 + £48 + £8 + £2 + £2 + £2 = £310.

So, in an attempt to recover or original loss of £8 and achieve the target of winning £2 per bet, we will have lost a further £248 on top of the previous losses of £56.

At this point, it is advised that betting should cease, unless the betting bank is so large that the losses can be disregarded, since the stake of the next bet will be £248 + £48 + £8 + £2 + £2 + £2 + £2 = £312 and the liability will be £1,248, assuming odds of 5.0.

A slightly less aggressive variant of the above is to, following a loss, recover the losses and disregard the £2 per bet profit element.

If this method is used, the worse case scenario is as follows:

Suppose that the first selection of the day is layed to lose at odds of 5.0 (4/1).

The stake, in order to win £2, must be set to £2.

The liability on the bet, should the selection win, is, therefore, £2 x (5 - 1) = £2 x 4 = £8.

If the selection loses and we win our bet, we have achieved the objective of winning £2 per bet and we simply move on to the next bet.

The stake on the next bet will be the usual £2.

If, on the other hand, the selection wins and the bet is lost, we will incur an £8 loss.

Now, let us consider the second bet, given that the previous bet lost:

Suppose that the second selection of the day is layed to lose at odds of 5.0 (4/1).

On the first bet, we lost £8.

In order to recover our losses on the next bet, the stake must be set to £8. This is four times our normal bet! If the selection wins, and we lose our bet, we will lose £8 x (5 - 1) = £8 x 4 = £32.

If the second selection loses and we win our bet, we will win £8.

We will have then recovered the losses of £8 on the first bet and will have achieved our objective.

We will therefore simply move on to the next bet. The stake on the next bet will be the usual £2.

If, on the other hand, the second selection wins and the bet is lost, we will have failed in the objective to recover the lost £8.

In addition, we will have lost a further £32. We will have therefore failed to meet the objective by the sum of £32 + £8 = £40.

So, in an attempt to recover losses of £8, we have lost a further £32.

Now, let us consider the third bet, given that the previous two bets both lost:

Suppose that the third selection of the day is layed to lose at odds of 5.0 (4/1).

On our first bet, we lost £8.

On the second bet, we lost £32.

Therefore, the stake on our third bet becomes £8 + £32 = £40. This is twenty times the normal bet!

If the selection wins, and we lose our bet, we will lose £40 x (5 - 1) = £40 x 4 = £160.

If the third selection loses and we win our bet, we will win £40.

We will have recovered our losses of £8 on the first bet and £32 on the second bet and will have achieved our objective.

We will therefore simply move on to the next bet.

The stake on the next bet will be the usual £2.

If, on the other hand, the third selection wins and the bet is lost, we will have failed in our objective to win back the lost £40.

In addition, we will have lost a further £160.

We will have therefore failed to meet our objective by the sum of £160 + £32 + £8 = £200.

So, in an attempt to recover the original losses of £8 and achieve the target, we will have lost a further £160 on top of our previous losses of £40.

The advantage of this variant is that, following three consecutive losses, £200 would be lost instead of the £304 loss using the more aggressive strategy.

From the above, it can be seen that as a losing run increases, the stake grows exponentially.

Now, let us consider the system from a backer’s point of view:

Suppose that the aim is to obtain a profit of £5 per bet.

For the moment, we will ignore the betting exchange’s commission on all winning bets.

Suppose that the first selection is backed to win at odds of 7.0 (6/1).

The stake, in order to win £5, must be set to £2 (the minimum bet on betting exchanges).

The liability on the bet, should the selection lose, is £2.

Should the selection win, we would win £2 x (7 - 1) = £2 x 6 = £12 and exceed our £5 target.

We would then carry on as normal and set the stake on the next but such that we would win £5 if the selection wins.

If the first selection loses and the bet is lost, we would lose £2.

The target win on the next bet, therefore, becomes £2 (loss from the previous bet) + £5 (to be won on the previous bet) + £5 (to be won on the next bet) = £12.

Suppose that the odds on the second selection are 3.0.

The stake is calculated by dividing the target profit (£12) by the odds minus 1.0 = 12/(3 -1) = 12/2.0 = 6.0.

The stake is therefore £6.00.

The liability on the bet, should the selection lose, is therefore £6.00.

If the selection wins, we would win £6.00 x (3.0 - 1.0) = £6.00 x 2.0 = £12 and the target would be met.

If the second selection loses and the bet is lost, we would lose £6.

The stake on the next bet, therefore, becomes: £6.00 (the losses from the second bet) + £2.00 (the losses from the first bet) + £5 (not won on the first bet) + £5 (not won on the second bet) + £5 (to be won on the third bet) = £23.00.

Suppose that the odds on the third selection are 3.0.

The stake is calculated bydividing the target profit (£23.00) by the odds minus 1.0 = £23/(3 - 1) = £23/2 = £11.5.

The stake is therefore £11.50. This is almost six times the original stake.

The liability on the bet, should the selection lose, is therefore £11.50.

Should the selection win, we would win £11.50 x (3.0 - 1.0) = £23.00 and the target would be met.

If the selection, on which the bet is placed, loses and the bet lost, we would lose £11.50.

The stake on the fourth bet therefore becomes: £11.50 (the losses from the third bet) £6.00 (the losses from the second bet) + £2.00 (the losses from the first bet) + £5 (not won on the first bet) + £5 (not won on the second bet) + £5 (not won on the third bet) + £5 (to be won on the fourth bet) = £39.50.

This is almost twenty times the original £2 stake.

When a winning bet is placed, a £5 profit (less commission) will be made.

At this point, the stake should be reduced, once again, to obtain a £5 profit.

At this point, it can be seen that the stakes, although still relatively small, are increasing rapidly.

It can also be seen that the size of the stakes when backing horses to win, do not increase as rapidly as they do when laying selections to lose.

However, losing runs, when backing horses to win, tend to be far larger than those associated with laying horses to lose.

To use this system, a large betting bank is required in case an extended run of losing bets is encountered.

One thing worthy of note: In the above examples, for the sake of simplicity, the commission levied by betting exchanges on winning bets have been completely ignored.

If the commission on winning bets were to be taken into account, the situation would be even worse.

The above is colloquially known as ‘chasing losses’.

It is definitely NOT to be recommended since even substantial betting banks can be completely wiped out in a relatively short period of time.

That is the end of the theoretical implications of loss recovery systems.

Here are two actual examples of what can happen when things go badly wrong:

The first example occurred early in 2006 when I was asked to review a Blackjack system for a website that sells gambling products, prior to its intended sale.

The literature, relating to the system, was extremely well written.

It was clear, concise and easy to follow. The system that it described was extremely simple to use.

It involved laying certain types of Black Jack hands under certain conditions.

I was advised that the strike rate of the system was in the region of 83%.

Whenever a losing bet was encountered, the system required that the stake on the next qualifying bet was increased to a level that would recover the funds lost on the previous bet.

If the subsequent bet was lost, the system required that the stake on the next qualifying bet was again increased to a level that would recover the funds lost on the pervious two bets.

The system documentation stated that, although on occasions, two consecutive losing bets had been encountered, three consecutive losing bets had never been encountered in the nine months that the system had been used by its creator.

The creator even supplied his Betfair P/L records as proof of his claims.

The system’s literature also stated that whenever the betting bank had increased to twice its initial value, 50% should be withdrawn and used to create an emergency betting bank.

The system’s literature also advised that a daily profit target should be set and, once achieved, betting for the day should cease.

I decided to test the system for two hours per day over a period of one week.

On the first two days, the system behaved flawlessly and the maximum number of consecutive losing bets encountered was one.

On day three, following perhaps 30 minutes of testing, the system encountered something that it had supposedly never encountered in the previous nine months - 3 consecutive losing bets.

Had I been placing real, instead of theoretical, bets, I would have lost the whole of my betting bank in a matter of a few minutes.

I continued testing and less than one hour later, I encountered another 3 consecutive losing bets.

Again, had I been placing real, instead of theoretical, bets, I would have lost the whole of my second betting bank in a matter of a few minutes.

That was two betting banks that I would have lost in the space of one hour.

So, something that hadn’t occurred in the previous nine months occurred twice in the space of one hour.

Lady luck can indeed be a cruel mistress.

Given that the system’s strike rate was 83%, there was a 50% chance that it would encounter 4 consecutive losing bets and a 25% chance that it would encounter 5 consecutive losing bets.

I also calculated that if the system was used to place 1,000 bets, the maximum losing run that it would encounter was 4.

I ceased testing the system at this point and advised the web site accordingly.

The second example occurred in mid-2006 when I was approached by a betting club and asked if I would like to become a member.

The club was run by a commercial organisation. It placed lay bets, on behalf of its members, using a system that, according to its creator, was bullet-proof.

Each member of the club contributed to a fund which formed the betting bank.

The past performance of the system over the past year had been impressive to say the least.

I contacted the organisation and requested additional information prior to joining.

Given that the information was not forthcoming, I decided to decline their offer.

One year later, I received a description of the system, anonymously.

I will spare you the details of the system for fear that someone may use it.

I have no wish to inflict it on a largely unsuspecting world because, although it sounds extremely plausible, it is also deeply flawed.

To make matters even worse, whenever a losing bet was encountered, the system required that the stake on the next qualifying bet was increased to a level that would recover the funds lost on the previous bet.

If the subsequent bet lost, the stake on the next qualifying bet was again increased to a level that would recover the funds lost on the previous two bets.

The system documentation stated that, although on occasions, two consecutive losing bets had been encountered, three consecutive losing bets had never been encountered in the two years that the system had been operating. Does this sound familiar?

In early December 2007, the system encountered something that it had never encountered before in the two years that it had been operating - 3 consecutive losing bets.

On that one day, most of the betting bank was lost.

Directors of the organisation were able to secure some additional funding by taking out loans, selling their cars and re-mortgaging their homes. Such was their (misguided) belief in the system.

The funds were used to replenish the betting bank.

During the remainder of December 2007, and for the vast majority of January 2008, the system behaved impeccably and began, once again, to generate profits.

The betting bank began to slowly recover.

Then, on 30th. January 2008, five horses were layed. Four of them won, three of which were consecutive. What little remained of the betting bank was also lost.

In March 2008, I learned that the organisation had gone into receivership and had reported debts of over £300,000.

Lightning struck twice in less than eight weeks - or so they thought.

In fact, it wasn’t surprising that lightning had struck twice. What was surprising was that it hadn’t even struck once in the previous two years.

The fact of the matter was that the basic logic behind both the system and the staking plan was seriously flawed.

It wasn’t a case of - would the system fail? it was more a case of - when would the system fail?

For such a system to last two years without a major catastrophe was a staggering achievement, but it wasn’t down to the management of the system, it was down to one thing and one thing only - GOOD LUCK.

Eventually, the run of good luck had to come to an end, and end it did. TWICE within less than eight weeks.

As for the members of the betting club - they lost every penny that they had invested. In some cases, this amounted to thousands of pounds.

We hope that you now understand why it isn’t a good idea to chase your losses.