Before we begin this article, it should be stated that the words Trading and Arbitraging are interchangeable. They are, in fact, different words for exactly the same thing.

One dictionary defines Arbitraging as ‘the simultaneous buying and selling of assets in different markets or in derivative forms, taking advantage of the differing prices.’

A second defines it as the trading of one currency for another in the hope of taking advantage of small differences in the currency conversion rates in order to achieve a profit.

For example, if a currency trader decides to trade US Dollars for Sterling and then Sterling for Euro and then Euro for US Dollars and if $1.00 buys £0.7 and if £1 buys 9.5 Euros and if 1 Euro buys $0.16, then an arbitrage trader who starts with $1.00 earns: 1 x 0.7 x 9.5 x 0.16 = $1.064.

This produces a profit of 6.4 cents or 6.4%. 6.4 cents doesn’t seem much but if, instead of trading $1, the trader begins with $1,000,000, the profit would be $64,000! This is not a bad day’s profit, to say the least.

From the above, it can be seen that Arbitrage isn’t about buying or selling anything. It’s about deciding whether the price of an item will increase or decrease and then taking advantage of the price movement.

A currency trader doesn’t buy or sell a currency with the intention of keeping it. He only buys a currency to sell it on at a profit.

So, what has all this talk of buying and selling of currency got to do with betting on horses?

Well, when a bet is placed on a horse to win, what is actually happening is that the right to a profit, if the horse obliges and wins, is being purchased.

When a horse is layed to lose, what is actually happening is that the right to a profit, if the horse obliges and loses, is being purchased.

Before the end of the race, the bet can be sold for a higher price than it was bought. In this way, a guaranteed profit will be made, regardless of the outcome of the race.

The best way to described arbitraging is by way of the following example:

It is 3 pm. A horse called ‘Arbit’ is running in the 3:30 race at Southdown.

The odds on Arbit on one of the betting exchanges is currently 5.0.

At 2:30 pm, the odds on Arbit started to slowly increase. It, therefore, looks probable that the odds on Arbit will continue to increase.

A £10 bet is therefore placed on Arbit to lose.

It is now 3:15. The odds on Arbit have increased to 10.0.

The second bet of £5 is therefore placed on Arbit, but this time to win.

Now let’s look at the maths:

Bet 1 (3:00 pm):

If Arbit wins the race, the loss will be £10 x (5.0 - 1) = £10 x 4 = £40.

If Arbit loses the race, the profit will be £10 (the stake).

Bet 2 (3:15pm)

If Arbit wins the race, the profit will be £5 x (10.0 - 1) = £5 x 9 = £45.

If Arbit loses the race, the loss will be £5 (the stake).

Now let’s look at the net effect of the two bets:

Arbit wins the race:

Loss on bet 1 = £40.

Profit on bet 2 = £45.

Net profit = £45 - £40 = £5.

Arbit loses the race:

Profit on bet 1 = £10 (the stake).

Loss on bet 2 = £5 (the stake).

Net profit = £10 - £5 = £5.

From the above, it can be seen that, regardless of whether Arbit wins or loses the race, a profit of £5 will be made.

What has actually been achieved by placing the two bets?

A bet was ‘purchased’ on Arbit at a particular price and re-sold at a higher price, thus making a profit in the process.

What actually happened was that a bet was purchased on Arbit to win for £5 (the back bet) and then a bet was sold on Arbit to lose for £10 (the lay bet).

The result was a £5 profit.

The result of the race was immaterial.

This is arbitraging or trading.

We have now dealt with the case where we anticipate that a horse’s odds may well increase.

Now let’s deal with the case where we think that a horse’s odds will decrease.

It is 5pm. A horse called ‘Tradeit’ is running in the 5:30 race at Southdown.

The odds on Tradeit on one of the betting exchanges is currently 8.0.

At 4:30pm, the odds on Tradeit began to slowly decrease. It therefore looks probable that the odds on Tradeit will continue to decrease.

A £5 bet is therefore placed on Tradeit to win.

At 5:15, the odds on Tradeit have decreased to 4.0.

A second bet of £10 is therefore placed on Tradeit, this time to lose.

Now let’s look at the maths:

Bet 1 (5:00pm):

If Tradeit wins the race, the profit will be £5 x (8.0 - 1) = £5 x 7 = £35.

If Tradeit loses the race, the loss will be £5 (the stake).

Bet 2 (5:15pm):

If Tradeit wins the race, the loss will be £10 x (4.0 - 1) = £10 x 3 = £30.

If Tradeit loses the race, the profit will be £10 (the stake). Now let’s look at the net effect of the two bets:

Tradeit wins the race:

Profit on bet 1 = £35.

Loss on bet 2 = £30.

Net profit = £35 - £30 = £5.

Tradeit loses the race:

Loss on bet 1 = £5 (the stake).

Profit on bet 2 = £10 (the stake).

Net profit = £10 - £5 = £5.

From the above, it can be seen that, regardless of whether Tradeit wins or loses the race, a profit of £5 will be made.

What has actually been achieved by the placing the two bets?

A bet was ‘purchased’ on Tradeit at a particular price and then sold at a higher price, thus making a profit in the process.

What actually happened was that a bet was purchased on Tradeit to win for £5 (the back bet) and then a bet was sold on Tradeit to lose for £10 (the lay bet).

The result was a £5 profit.

The result of the race was immaterial.

This is what Arbitraging is all about.

It isn’t about betting on the outcome of a race.

It’s about profiting from a movement of the odds on a horse, rather than on a horse’s performance in a race.

That is why the outcome of a race is unimportant in arbitraging.

What is important in arbitraging is that the odds on a horse move in the predicted direction.

The more that the odds move in the predicted direction and the greater the size of the bet, the larger will be the profit.

Given the above, here’s a good question:

If arbitraging a bet always leads to a profit, regardless of the outcome of the race, why shouldn’t all bets be arbitraged?

And here are the reasons for not arbitraging all bets:

Firstly, the odds on the selected horse might remain fairly constant.

As a result, arbitraging a bet may not be possible due to the lack of movement in the odds and the initial bet may need to be honoured.

Secondly, the odds on the horse may move in the opposite direction to the one anticipated.

As a result, arbitraging a bet may not be possible and the initial bet may need to be honoured.

Thirdly, if a bet is arbitraged, there is a penalty to be paid because we can’t have something for nothing, not in this world and certainly not in horse racing!

The penalty that must be paid is that most of the profit potential on a horse must be sacrificed in order to secure a smaller arbitrage gain.

Again, this is best explained by way of an example.

Let us suppose that ‘Arbit’ is running in the 3:30 race at Southdown.

Let us further suppose that the odds on Arbit on one of the betting exchanges is 5.0 and a £5 bet is placed on Arbit to win.

The odds on Arbit then fall to 2.5.

A second (lay) bet of £10 is then placed on Arbit, this time to lose.

Now let’s look at the maths:

Bet 1

If Arbit wins the race, the profit will be £5 x (5.0 - 1) = £5 x 4 = £25.

If Arbit loses the race, the loss will be £5 (the stake).

Bet 2

If Arbit wins the race, the loss will be £10 x (2.5 - 1) = £10 x 1.5 = £15

If Arbit loses the race, the profit will be £10 (the stake).

Now let’s look at the net effect of the two bets:

Arbit wins the race:

Profit on bet 1 = £25.

Loss on bet 2 = £10.

Net profit = £25 - £10 = £15.

Arbit loses the race:

Loss on bet 1 = £5 (the stake)

Profit on bet 2 = £10 (the stake).

Net profit = £10 - £5 = £5.

From the above, it can be seen that if Arbit wins the race, the profit will be £10 and if Arbit loses the race, the profit will be £5.

The situation is such that, irrespective of Arbit’s performance in the race, a profit will be made.

Now let’s change the situation slightly.

Let us suppose that a £5 bet is placed on Arbit to win at odds of 5.0 but a second lay bet is not placed when the odds on the horse fall.

In other words, the bet is not arbitraged.

If Arbit loses the race, £5 will be lost. If, however, Arbit wins the race, £20 will be won.

Had the bet been arbitraged, a potential £20 win or £5 loss would have been replaced by a certain £5 win.

This is a case of ‘is a bird in the hand worth two in a bush?’.

Is a certain £5 win better than a potential win of £20?.

Only you can answer this question.

Ok, so now we know how to create a winning bet regardless of the outcome of a race.

Now let’s look at how to create a free bet.

There are two ways of achieving this.

Here’s the first:

From the above examples, we can see that a profit can be made, regardless of whether or not the horse wins the race, provided that the attempted arbitrage succeeds.

We now simply place the amount won on another horse, either in the same race or in a different race.

If the bet loses, nothing will be lost since all we will lose is the amount that was won on the previous bet.

However, if the second bet wins, we can look forward to adding to the amount that was won on the first bet through the successful arbitraging of our selection.

And here’s the second (which is a little more complex):

To explain this method, we need to go back to the 3:30 race at Southdown and our old friend Arbit.

Let us suppose that the odds on Arbit at 3 pm are 5.0 and that a £5 bet is placed on Arbit to lose the race.

The odds on Arbit then begin to increase until, at 3:15, they reach 11.0.

At this point, a £2 bet is placed on Arbit to win the race.

Now let us look at the maths:

Bet 1 (3:00 pm):

If Arbit wins the race, the liability is £5 x (5.0 - 1) = £5 x 4 = £20.

If Arbit loses the race, we will win £5 (the stake).

Bet 2 (3:15 pm):

If Arbit wins the race, the profit would be £2 x (11.0 - 1) = £2 x 10 = £20.

If Arbit loses the race, the liability is £2 (the stake).

Now let’s look at the net effect of the two bets based upon the race outcome:

Arbit wins the race:

Loss on bet 1 = £20.

Profit on bet 2 = £20.

Net profit = £20 - £20 = £0.

Arbit loses the race:

Profit on bet 1 = £5 (the stake)

Loss on bet 2 = £2 (the stake).

Net profit = £5 - £2 = £3.

What have we achieved in placing the two bets?

The second bet nullified the liability of the first bet at a cost of £3 less than that of the initial bet.

As a result, if Arbit wins the race, we will break even.

If Arbit loses the race, we will win £3.

What we have actually achieved in placing the two bets is to totally nullify the initial lay bet and leave ourselves with a potential profit of £3 that will be won if Arbit loses the race.

In other words, we have created a free bet on Arbit.

We will break even if Arbit wins and will make a £3 profit if Arbit loses.

We have seen how to arbitrage a selection in order to secure a guaranteed profit, regardless of the outcome of a race, provided that its odds change.

We have also seen how to arbitrage a selection in order to create a free bet.

Is there anything else that we can achieve by arbitraging our selections?

Well, yes there is.

Let us suppose that you have:

• Never placed a bet on a horse before • Never used a betting exchange before • A new system that requires further testing before it can be considered fully tested

In such cases, the cautious investor may wish to stake less than the minimum allowed by the betting exchanges in order to reduce the liability of bets at a time when a system is at its most vulnerable.

On some betting exchanges, the minimum stake is £2 per bet.

Note that the minimum is applied to the stake and not to the liability on the bet.

Also note that when laying a horse to lose, the liability is a multiple of the stake of the bet.

In order to place a bet on a betting exchange such that the stake is less than the minimum allowed, a form of arbitraging can be used.

By way of an example, let us suppose that we lay a horse to lose using a stake of £3 and then back it to win for £2.

What has this achieved?

Effectively, a horse has been layed to lose using a stake of £1 (£3 - £2).

This is £1 less than the minimum £2 stake allowed by the betting exchange.

As a result, the liability on the bet is less than it would otherwise be had the minimum stake been used.

To illustrate this point, let us look at the following example:

Let us suppose that the odds on a horse on a betting exchange are 3.0.

Let us also suppose that the horse is layed to lose using a stake of £3 and then backed to win using a stake of £2.

Now let’s look at the maths:

Bet 1

If the horse wins the race, the liability would be £3 x (3.0 - 1) = £3 x 2 = £6.If the horse loses the race, we will win £3. Bet 2

If the horse wins the race, the profit would be £2 x (3 - 1) = £2 x 2 = £4.

If the horse loses the race, we will lose £2.

Now let’s look at the net effect of the two bets based upon the race outcome:

Horse wins the race:

Loss on bet 1 = £6.

Profit on bet 2 = £4.

Net profit = £4 - £6 = -£2.

Horse loses the race:

Profit on bet 1 = £3 (the stake)

Loss on bet 2 = £2 (the stake).

Net profit = £3 - £2 = £1.

What have we achieved in placing the two bets?

If the horse is layed to lose using the minimum £2 stake allowed by the betting exchange and it won, the liability would be £6.

By placing the second back to win bet, the liability, should the horse win, has been reduced to £2.

Throughout this article, we have referred to the arbitraging of one selection per race.

There is nothing, however, to prevent multiple selections per race from being arbitraged.

It is even possible to arbitrage the same selection multiple times.

Here is an example of multiple arbitrages

A friend of yours sends you a text message. A friend of his is a stable lad.

The stable lad has told him about three particular horses that are running in the 3:30 at Southdown this afternoon.

The first horse is called Purple Patch. It did some exceptional work on a training gallop two days ago. The horse is fit, well and raring to go. If it doesn’t win, it will come close. It is definitely worth an ‘each way’ bet.

Its odds on the betting exchanges, at the moment, are 10.0. It is the fifth favourite.

The second horse is called Bank Balance. Although this horse won its last race ten days ago by a country mile and at a canter, it has not been running well in training of late and may benefit from a rest. Its last race was a close run affair and, apparently, took more out of the horse than was originally thought.

Bank Balance is therefore unlikely to run well. It is currently the race favourite at odds of 2.5 on the betting exchanges. Its favouritism is purely based on its performance in its last race.

The third horse is called Dun ‘N Dusted. In his last race, Dun ‘N Dusted came a close second to a very well regarded horse who has since gone on to record a second win. That race took a lot out of Dun ‘N Dusted and has since lost weight. The horse would probably benefit from a rest.

As a result, it is very likely that Dun ‘N Dusted will finish close to the back of the field. It is currently the second favourite in the betting at odds of 4.0 on the betting exchanges. This is largely based on its last performance.

Your friend tells you that the information about these three horses is not widely known at the moment. As such, their odds do not fully reflect this information that he has given you.

Based on your friend’s comments, you decide that when the information that he gave you becomes widely known and knowing your friend - it most likely will, the odds on Purple Patch will decrease and the odds on Bank Balance and Dun ‘N Dusted will increase.

You, therefore, place the following bets on a betting exchange:

Bet 1: £5 win bet on Purple Batch at odds of 10.0.

Bet 2: £10 lay bet on Bank Balance at odds of 2.5.

Bet 3: £10 lay bet on Dun ‘N Dusted at odds of 4.0.

You then text all of your friends and give them the information, knowing that most of them will pass on the information to all of their friends.

One hour later, you visit the betting exchange again.

You notice that the odds on Purple Patch have fallen, as was expected, from 10.0 to 3.0 and it is now favourite.

The odds on Bank Balance have, as was also expected, increased from 2.5 to 9.0 and the odds on Dun ‘N Dusted are now 10.0, from 4.0.

You, therefore, place the following bets on a betting exchange:

Bet 4: £10 lay bet on Purple Batch at odds of 3.0.

Bet 5: £5 win bet on Bank Balance at odds of 9.0.

Bet 6: £5 win bet on Dun ‘N Dusted at odds of 10.0.

Now let’s look at the net effect of the six bets:

Purple Patch wins:

Profit on bet 1 = £5 x (10.0 - 1) = £5 x 9 = £45.

Loss on bet 4 = £10 x (3.0 - 1) = £10 x 2 = £20.

Net profit = £45 - £20 = £25.

Purple Patch loses:

Loss on bet 1 = £5 (the stake).

Profit on bet 4 = £10 (the stake).

Net profit = £10 - £5 = £5.

Bank Balance wins:

Loss on bet 2 = £10 x (2.5 - 1) = £10 x 1.5 = £15.Profit on bet 5 = £5 x (9.0 - 1) = £5 x 8 = £40.

Net profit = £40 - £15 = £25.

Bank Balance loses:

Profit on bet 2 = £10 (the stake).

Loss on bet 5 = £5 (the stake).

Net profit = £10 - £5 = £5.

Dun ‘N Dusted wins:

Loss on bet 3 = £10 x (4.0 - 1) = £10 x 3.0 = £30.

Profit on bet 6 = £5 x (10.0 - 1) = £5 x 9 = £45.

Net profit = £45 - £30 = £15.

Dun ‘N Dusted loses:

Profit on bet 3 = £10 (the stake).

Loss on bet 6 = £5 (the stake).

Net profit = £10 - £5 = £5.

The net profit on the six bets depends on which horse wins.

If Purple Patch wins, the profit = £25 + £5 (because Bank Balance lost) + £5 (because Dun ‘N Dusted lost) = £35.

If Bank Balance wins, the profit = £25 + £5 (because Purple Patch lost) + £5 (because Dun ‘N Dusted lost) = £35.

If Dun ‘N Dusted wins, the profit = £15 + £5 (because Purple Patch lost) + £5 (because Bank Balance lost) = £25.

If any other horse wins, the profit = £5 (because Bank Balance lost) + £5 (because Purple Patch lost) + £5 (because Dun ‘N Dusted lost) = £15.

Notice that the arbitraging profit is cumulative in that the net profit is the sum of the arbitraging profits made on the three individual horses concerned.

So, no matter what the outcome of the race is, a maximum of £35 and a minimum of £15 will be won, depending on which horse wins the race.

Now, let’s move on.

Not long after the bets have been placed, you receive another text from your friend.

The text says that he has received a tip from a stable lad at another yard that a horse called ‘Getting Better’ will win the race in which Purple Patch, Bank Balance and Dun ‘N Dusted are also running.

When Getting Better last ran, his performance was well below par. Hence, the horses' current odds of 12.0 on the betting exchanges.

A visit from the vet, following the race, revealed that Getting Better was suffering from a virus. The horse is now completely virus-free and has been putting in some stunning performances on the gallops lately.

This information is known to few people. Hence, the horse’s odds.

From this latest information, you conclude that Getting Better’s odds will fall dramatically when the information begins to circulate and that the odds on Purple Patch, the current favourite, will increase.

You, therefore, place the following bets on a betting exchange:

Bet 7: £5 win bet on Getting Better at odds of 12.0.

Bet 8: £10 lay bet on Purple Patch at odds of 3.0.

It is now 5 minutes before the start of the race. The odds on Getting Better have fallen to 2.0 and the odds on Purple Patch have increased to 6.0.

You, therefore, place the following bets on a betting exchange:

Bet 9: £10 lay bet on Getting Better at odds of 2.0.

Bet 10: £5 win bet on Purple Patch at odds of 6.0.

Now let’s look at the net effect of the latest four bets and the profit made on each of the two horses:

Getting Better wins:

Profit on bet 7 = £5 x (12.0 - 1) = £5 x 11 = £55.

Loss on bet 9 = £10 x (2.0 - 1) = £10 x 1 = £10.

Net profit = £55 - £10 = £45.

Getting Better loses:

Loss on bet 7 = £5 (the stake).

Profit on bet 10 = £10 (the stake).

Net profit = £10 - £5 = £5.

Purple Patch wins:

Loss on bet 8 = £10 x (3.0 - 1) = £10 x 2 = £20.

Profit on bet 10 = £5 x (6.0 - 1) = £5 x 5 = £25.

Net profit = £25 - £20 = £5.

Purple Patch loses:

Profit on bet 8 = £10 (the stake).

Loss on bet 10 = £5 (the stake).

Net profit = £10 - £5 = £5.

Now let’s look at the overall effect of the 10 bets.

Again, the amount of profit depends upon which horse wins the race.

If Purple Patch wins, the profit = £25 + £5 (because Bank Balance lost) + £5 (because Dun ‘N Dusted lost) + £5 (because Getting Better lost) = £40.

If Bank Balance wins, the profit = £25 + £5 (because Purple Patch lost) + £5 (because Dun ‘N Dusted lost) + £5 (because Getting Better lost) = £40.

If Dun ‘N Dusted wins, the profit = £15 + £5 (because Purple Patch lost) + £5 (because Bank Balance lost) + £5 (because Getting Better lost) = £30.

If Getting Better wins, the profit = £45 + £5 (because Purple Patch lost) + £5 (because Bank Balance lost) + £5 (because Dun ‘N Dusted lost) = £60.

If any other horse wins, the profit = £5 (because Bank Balance lost) + £5 (because Purple Patch lost) + £5 (because Dun ‘N Dusted lost) + £5 (because Getting Better lost) = £20.

Notice again that the arbitraging profit is cumulative in that the overall profit is the sum of the arbitraging profits made on the individual horses concerned.

So, no matter what the race outcome is, a maximum of £60 and a minimum of £20 will be won, depending upon which horse wins the race.

Selections should be arbitraged when the odds have increased to the point where the arbitrage profit requirements have been satisfied, rather than waiting for the odds to increase, or decrease, still further. That way, the profit is secure.

A point worthy of note is that the odds required to effect the required arbitrage profit may exist on a betting exchange for a few seconds only and may disappear just as quickly as they appeared.

To avoid the disappointment of missing the required odds, it is advised that the second bet is created as soon after the first bet has been submitted for matching to the betting exchange.

The second bet should contain the stake and the odds required.

The bet should be submitted to the betting exchange for matching as soon as the first bet has been matched – and not until.

It will remain on the exchange, in an unmatched state, until such time as the required odds become available.

If they do, the bet will be matched and the required arbitrage profit will have been secured.

Be aware, however, that if a horse, running in the same race, is withdrawn at any time following the placing of the second (unmatched) bet, or the second bet has not been matched by the start of the race, the bet will be cancelled by the betting exchange.

If this happens, simply regenerate the bet and place it on the system to await matching.

As has now become apparent, arbitraging involves the placing of a win and a lay bet on the same horse.

Traditional bookmakers do not allow members of the public to place lay bets on horses.

As a result, the lay bet must be placed on one of the betting exchanges.

The win bet, however, can be placed either on one of the betting exchanges or with a traditional bookmaker.

The bet can even be placed with a traditional bookmaker using the internet or by phone.

Therefore, it is possible to sign onto a betting exchange using one internet session and to sign onto a traditional bookmaker using a second internet session and monitor the odds in both systems with a view to taking advantage of the odds differential between the two systems in order to arbitrage a bet.

We have now seen that in order to arbitrage a selection, we must first identify a selection whose odds, we feel, will either increase or decrease.

Having identified a selection, we must place the initial bet.

If we feel that the selection’s odds will increase, we must place the lay bet first.

When the odds have increased, we must then place a second bet on the selection. In this case, the second bet must be for the selection to win.

If we feel that the selection’s odds will decrease, we must place the back to win bet first.

When the odds have decreased, we must then place a second bet on the selection.

In this case, the second bet must be for the selection to lose. It should also be noted that the first bet if placed on a betting exchange, must be matched before the second bet is placed. Otherwise, the second bet may get matched and the first bet may not. In this case, we may not be presented with a chance to arbitrage the bet.

In order to benefit financially from arbitraging, the potential profit on the first bet must be greater than the potential loss on the second bet and the potential profit on the second bet must be greater than the potential loss on the first bet.

The odds movement on a selection and the stakes placed on the selection must be such that the above is true. Otherwise, any attempt to arbitrage a selection will result in a potential loss.

Here are two examples which illustrate this point:

Example 1:

Let us suppose that the odds on a selection on one of the betting exchanges is 5.0 and a £5 bet is placed on it to win. The odds then fall to 2.5 and a second (lay) bet of £10 is placed on the selection to lose.

Now let’s look at the maths:

Bet 1

If the selection wins, the profit will be £5 x (5.0 - 1) = £5 x 4 = £25.

If the selection loses, the loss will be £5 (the stake).

Bet 2

If the selection wins, the loss will be £10 x (2.5 - 1) = £10 x 1.5 = £15

If the selection loses, the profit will be £10 (the stake).

The potential profit on bet 1 (£25) is greater than the potential loss on bet 2 (£15) and the potential profit on bet 2 (£10) is greater than the potential loss on bet 1 (£5).

As such, we have effected an arbitrage where we are in a winning position irrespective of the outcome of the race.

Example 2:

Let us suppose that the odds on a selection on one of the betting exchanges is 5.0 and a £5 bet is placed on it to win.

The odds then increase to 6.0 and a second (lay)bet of £10 is placed on the selection to lose.

Now let’s look at the maths:

Bet 1

If the selection wins, the profit will be £5 x (5.0 - 1) = £5 x 4 = £25.

If the selection loses, the loss will be £5 (the stake).

Bet 2

If the selection wins, the loss will be £10 x (6.0 - 1) = £10 x 5.0 = £50

If the selection loses, the profit will be £10 (the stake).

Although the potential profit on bet 2 (£10) is greater than the potential loss on bet 1, the potential profit on bet 1 (£25) is less than the potential loss on bet 2 (£50) and (£5).

As such, we will win (£10 - £5 = £5) if the horse loses but will lose (£50 - £25 = £25) if the horse wins.

Therefore, we have failed to arbitrage the selection because we will not win irrespective of the outcome of the race.

Betting exchanges charge a commission on all winning bets. Therefore, when using a betting exchange, it should be remembered that the commission needs to be taken into account when considering arbitraging.

Now that the principles of arbitraging have been explained and examples have been provided, let’s move on.

Most laying selection systems are profitable at the traditional bookmaker’s Starting Prices (SP).

However, as we have already seen, a selection cannot be layed to lose using a traditional bookmaker since they do not accept lay bets.

As such, lay bets can only be placed on a betting exchange.

The odds to lay a selection on a betting exchange are often higher than SP.

On average, the odds on betting exchanges are approximately 20% higher than SP.

Therefore, potential losses are 20% higher, on average.

In addition, betting exchanges levy a commission on all winning bets.

When the betting exchange’s higher odds and the commission on winning bets are taken into account, few laying systems are profitable.

The situation is made worse if the laying system, or the tips from one, are purchased from a third party since these additional costs only serve to reduce profits even further.

One way to overcome the above issue is to lay selections at either SP or better, odds.

If this could be achieved, potential losses would be reduced and, thus, profits would be increased.

So, how do we lay a selection at either SP or better, odds on a betting exchange?

Given that the topic of this chapter relates to arbitraging, it is a fair assumption that the answer has something to do with arbitraging.

If you were to assume this, then you would be perfectly correct.

Arbitraging isn’t just about creating a winning situation regardless of the outcome of a race.

It is also about reducing the odds of a selection which is to be layed to lose and increasing the odds of a selection which is to be backed to win.

Thus far, all of the examples in this article relate to a type of arbitraging whose purpose it is to create a winning situation regardless of the outcome of a race.

This type is what I refer to as ‘fully’ arbitraging since the potential liability on the first bet is fully offset by the potential profit on the second bet and the potential liability on the second bet is fully offset by the potential profit on the first bet.

A second type of arbitraging exists, however. Its purpose is to reduce the odds of a selection which is to be layed to lose and to increase the odds of a selection which is to be backed to win.

The remainder of this article will concentrate on this topic which I refer to as ‘partially’ arbitraging.

Firstly, let’s consider reducing the odds of a selection that is to be layed to lose.

To illustrate this point, let us look at the following example:

Let us suppose that the odds of a selection on one of the betting exchanges is 4.0 and a £5 bet is placed on the selection to lose.

The odds on the selection then increase to 5.0 and a second bet of £3 is placed on the selection to win.

Now let’s look at the maths:

Bet 1

If the selection wins, the loss will be £5 x (4.0 - 1) = £5 x 3 = £15.

If the selection loses, the profit will be £5 (the stake).

Bet 2

If the selection wins, the profit will be £3 x (5 - 1) = £3 x 4 = £12.

If the selection loses, the loss will be £3 (the stake).

Now let’s look at the net effect of the two bets:

Our selection wins:

Loss on bet 1 = £15.

Profit on bet 2 = £12.

Net profit = £12 - £15 = -£3 (a loss).

Our selection loses:

Profit on bet 1 = £5 (the stake).

Loss on bet 2 = £3 (the stake).

Net profit = £5 - £3 = £2.

If we divide the net loss by the net win on the two bets, we get 3/2 = 1.5. Adding 1.0 to the result gives us 2.5.

These are the betting exchange odds that the selection was actually layed at, as opposed to the betting exchange odds of 4.0 that the selection would have layed at had the second bet not been placed on the selection to win.

As we can see from the above, the bet has only been ‘partially’ arbitraged because there is an outstanding liability on the two bets of £3.

Given that the selection is expected to lose, however, this is perfectly acceptable.

In addition, the profit, should the selection lose as expected, is greater than the profit that would have been achieved had the selection been fully arbitraged.

What has been achieved by placing the two bets on the selection?

Well, the liability of the bet has been reduced from £15 to £3.

In order to reduce the liability, it must be recognised however that a penalty had to be paid in that a reduction in the potential profit (from £5 to £2) had to be accepted.

Therefore, in order to reduce the liability of the bet, the potential profit had to be reduced - but to a lesser extent.

As a result, a selection has been layed to lose at odds that are less than those available on the betting exchange.

In fact, the selection has been layed at odds of 2.5 instead of 4.0.

This represents a reduction in odds of 100% x (4 - 2.5) / 4 = 100% x 1.5/4 = 25% x 1.5 = 37.5%.

We have therefore layed the selection to lose at odds which are 37.5% less than the odds available on the betting exchange.

This is likely to be less than the SP of the selection.

Of course, the odds of a selection will not always increase after it has been layed to lose.

But then, if it doesn’t, nothing will have been lost and, if it does, we will have gained.

We have now seen how to use partial arbitraging to reduce the odds at which a selection can be layed to lose.

Now, let’s consider increasing the odds of a selection that is to be backed to win on a betting exchange.

To illustrate this, let us look at the following example:

Let us suppose that the odds of a selection on one of the betting exchanges is 7.0 and a £5 bet is placed on the selection to win.

The odds on the selection then decrease to 3.0 and a second bet of £2 is placed on the selection to lose.

Now let’s look at the maths:

Bet 1

If the selection wins, the profit will be £5 x (7.0 - 1) = £5 x 6 = £30.

If the selection loses, the loss will be £5 (the stake).

Bet 2

If the selection wins, the loss will be £2 x (3.0 - 1) = £2 x 2 = £4

If the selection loses, the profit will be £2 (the stake).

Now let’s look at the net effect of the two bets:

Our selection wins:

Profit on bet 1 = £30.

Loss on bet 2 = £4.Net profit = £30 - £4 = £26.

Our selection loses:

Loss on bet 1 = £5 (the stake).

Profit on bet 2 = £2 (the stake).

Net profit = £2 - £5 = -£3.

If we divide the net win by the net loss on the two bets, we get 26/3 = 8.67.

Adding 1.0 to the result gives us 9.67.

These are the betting exchange odds that the selection was actually backed at, as opposed to the betting exchange odds of 7.0 that the selection would have backed at had the second bet not been placed on the selection to lose.

As we can see from the above, the bet has only been ‘partially’ arbitraged because there is an outstanding liability on the two bets of £3.

Given that the selection is expected to win, however, this is perfectly acceptable.

In addition, the profit, should the selection win as expected, is greater than the profit that would have been achieved had the selection been fully arbitraged.

What have we achieved by placing the two bets on the selection?

Well, we have reduced the liability on our bet from £5 to £3.

In order to reduce our liability, it must be recognised that we have had to pay a penalty in that we have been forced to accept a reduction in our potential profit from £30 to £26.

Therefore, in order to reduce the liability of the bet, the potential profit had to be reduced - but to a lesser extent.

As a result, a selection has been backed to win at odds that are greater than those available on the betting exchange.

In fact, the selection has been backed to win at odds of 9.67 instead of 7.0.

This represents an increase in odds of 100% x (9.67 - 7.0) / 7 = 100% x 2.67/7 = 267%/7 = 38.14%.

We have therefore layed the selection to lose at odds which are 38.14% greater than the odds available on the betting exchange.

This is likely to be much greater than the SP of the selection.

Of course, the odds of a selection will not always increase after it has been backed to win.

But then, if it doesn’t, we will have lost nothing and, if it does, we will have gained.

We have now seen how to use partial arbitraging to increase the odds at which a selection can be backed to win.

Arbitraging is such a relatively involved and a relatively complex area because it involves the placing of two or more coordinated bets.

As such, we have included a tips and hints section in this article.

Tips and Hints

• Before arbitraging activities begin, it is essential that the theory, methodology and implications are fully understood.

• Before arbitraging activities begin, it is essential that it is tested out on paper first.

• Prior to placing the initial bet, a plan should be created. The plan needs to cater for all circumstances such that, if and when they occur, a solution is at hand. This will avoid the necessity of having to make an on the spot decision under pressure, especially during in-running arbitraging i.e whilst the race is in progress. As arbitraging proficiency increases, the issues that arise and their possible solutions will become apparent.

• When first starting to arbitrage, it is advisable that minimum stakes are used. In this way, any misconceptions or errors will result in minimum losses.

• Learning to become a good arbitrager requires practice and patience. It is not a skill that can be developed overnight.

• Accept that mistakes are part of the learning process and that they will happen from time to time.

• If a decrease in a selection’s odds is anticipated, the win bet should be placed first and the lay bet placed when the odds decrease.

• If an increase in a selection’s odds is anticipated, the lay bet should be placed first and the back bet placed when the odds increase.

We have now learned what arbitraging is, what it can be used for and how to arbitrage a bet.

We have also learned how we can create a free bet and how to arbitrage multiple horses in the same race and even how to arbitrage the same horse multiple times.

We have also learned how we can reduce the odds of a selection that we would like to lay to lose and how to increase the odds of a selection that we would like to back to win.