How To Build A Betting System Portfolio That Survives

You finally get a system working. It has been tested, it makes a profit, and for a few weeks it feels like you have cracked it. Then it hits a bad losing run, and because it is the only thing you are running, the whole bank goes down with it. There is nothing to cushion the fall, and nothing to tell you whether this is normal variance (the ordinary ups and downs in results) or the system breaking.

That is the problem a betting system portfolio is built to ease. It can be the difference between an approach you can live with and one you drop at the worst moment.

By the end of this guide you will be able to do four things. Choose two or three genuinely different systems. Give each one a sensible return target. Check in your own records whether their bad patches overlap. Then run them from one shared bank with simple rules for when to leave them alone. It is a build, not a theory, so each chapter is a step.

In this guide

• What is a betting system portfolio?
• Why one system is not enough
• The three-system shape
• Set a different return target for each system
• Return means little without the number of bets
• Check that your systems really lose at different times
• Run every system from one shared bank
• Why a portfolio might survive when a single system would not
• Build it one tested system at a time
• Know when to leave it alone and when to pause a leg
• Common mistakes to avoid
• Frequently asked questions
• Do this today

What is a betting system portfolio?

A betting system portfolio is two or three different systems run together so their bad patches are less likely to land at the same time. The shape I use as an idea, not a tested result, is one system backing higher-priced horses, one backing lower-priced horses, and a laying system. Each gets its own return target, weighed against how many bets it gives you. Run from one shared bank, the combined results may be smoother than any single system on its own, but only if the systems really do lose at different times. That is something you check in your own records, not something the shape guarantees.

Think of it as the same idea as not putting every egg in one basket. The difference is that in betting you have to prove the baskets are separate. Two systems that look different on paper can still fail together if they lean on the same kind of race.

Why one system is not enough

A good system is not a smooth line of profit. It is a generally rising line with some ugly stretches in it. Methods tend to have a bad losing run sitting in them somewhere, and you will probably meet it. When it is the only thing you are running, that run is your whole experience of betting for as long as it lasts.

That is what breaks people. Not the long-run expectation, but the bank swings on the way there. A single system gives you no shelter from its own bad patch. The idea of a portfolio is that when one system is having its worst stretch, the others are less likely to be, so the bank notices less.

There is a second problem with one system. With nothing else to compare it against, you cannot tell a normal dry spell from a real fault. A portfolio gives you context. If one leg is cold while the others behave as usual, that is useful information. If all of them go cold together, that is a different question, and a better one to be asking.

The three-system shape

You do not need five systems. Two can do it, and three is plenty. What matters is that they are genuinely different, because two systems that bet on the same kind of horse in the same kind of race will tend to win and lose together, and that defeats the purpose. This shape is a rule of thumb, not something I have proved.

leg-higher-priced

The first system backs higher-priced horses. The idea is that value may be easier to find further out in the market, where there is more room for your judgement to matter. You accept a lower strike rate (how often it wins) in exchange for bigger-priced winners. It is worth capping the odds at a sensible ceiling rather than reaching for huge prices, because the winners dry up fast.

leg-lower-priced

The second system backs lower-priced horses. Its job is partly defensive. A portfolio that only backs outsiders is constantly opposing strong favourites. A system aimed at shorter prices wins more often and pays less when it does, which is the counterweight you want.

leg-laying

The third is a laying system. Laying means betting on a horse to lose, so it wins whenever the horse is beaten, which is most of the time. That gives a high strike rate and a steady run of small wins, which can fill in some of the gaps the backing systems leave. One honest caution: the liability on a big-priced lay (what you stand to lose if the horse wins) can be many times your stake. Keep the prices you lay sensible and the liabilities controlled, or the lay leg (the laying part of the portfolio) adds risk instead of calming things down.

verified-figures

Laying favourites across the board is not a proven approach. Here are the verified figures for three common signals. A/E means actual winners divided by the winners the odds implied, so 1.00 means results matched the prices. All figures are for GB and Ireland races from 1 July 2013 to 14 October 2026, Betfair SP, level one-point stakes, 2% commission, with fallers and pulled-up horses counted as losers:

• Betfair SP favourite: 167,257 races, wins 34.4%, A/E 1.00, return -1.4%, longest losing run 28.
• VDW Rank 1 (the horse our VDW rating system puts top in each race): 153,920 races, wins 21.6%, A/E 0.99, return -3.8%, longest losing run 59.
• Race Predictor top horse (the runner our model rates best in each race): 167,314 runners, wins 25.5%, A/E 1.01, return 0.0%, longest losing run 36.

These figures describe that stretch of racing and will not repeat exactly. What they suggest is that none of these signals, backed blindly, has stood clearly ahead of the market, which is why a lay leg needs its own narrower conditions and its own testing before it earns a place.

Set a different return target for each system

The instinct is to demand the same return from each system. That is the wrong way round. Each one does a different job in a different part of the market, so each can be held to a different standard.

As an idea, the high-priced backing system might be expected to earn the most per bet, because it carries bigger prices and longer losing runs. The lower-priced backing system might settle for less, because catching strong short runners at a fair price is worth having even at a slim margin. The lay leg will by its nature return the least per bet, because a lay risks more than it stands to win. I would be happy for a lay leg to sit near break-even after commission if it makes the rest of the portfolio steadier. That is a judgement call, not a tested rule.

None of this works unless the numbers underneath are real. Measure the edge on each system separately, against Betfair SP, before you trust it. A system you have not tested does not earn a place in the portfolio because the shape sounds right.

A simple way to record the targets is a short card for each system. Write down four things and keep them fixed:

• The job it does (for example, “backs higher-priced horses up to a set ceiling”).
• The return you need from it, after commission, to justify its place.
• How many bets you expect from it in a typical month.
• The losing run or drawdown at which you will look at it again (decided in advance).

Writing the card before you bet is the point. It stops you moving the goalposts when a leg has a bad month.

Return means little without the number of bets

A return on investment (ROI, profit as a share of what you staked) tells you half the story. The other half is how many bets the system gives you, because profit is the return multiplied by the money you put through it.

Picture a system with a lovely ROI that produces five bets a month. To earn anything meaningful you would have to stake heavily on each one. Now picture a more modest ROI with fifty or sixty bets a month. The smaller margin is spread across far more turnover, so the same profit comes from stakes you can actually manage. Often I would rather have the second system, because it is the one a bank can be built around. That is a preference, not a rule.

This is also why the systems naturally run at different frequencies. The lower the margin, the more bets you want to make the profit add up. The flip side is that more bets means you should stake them consistently from one shared bank. A portfolio only smooths the curve if each leg is sized against the same pot.

A quick check when you plan a leg: multiply your expected bets per month by your usual stake to get monthly turnover, then ask what return on that turnover you would need to feel the effort was worth it. If the answer is an unrealistic return, the leg does not have enough bets to matter, and you may need a wider filter or a different system.

Check that your systems really lose at different times

This is the step most people skip, and it is the one that decides whether you really have a portfolio. The aim is to see whether your systems have their bad spells at the same time.

overlap-method

1. Pick a fixed period, such as a week, and use the same period for every system.
2. For each system, record its result in units for every period, using your own tracked bets.
3. Mark each period as a losing period (below zero) or not.
4. Line the systems up side by side and count the periods where two or more systems lost together.
5. Compare that count with how often each system loses on its own.

If one system loses in most of the periods where another loses, they are leaning on the same thing, and they are not giving you much protection. As a rule of thumb (mine, not a proven threshold), if more than half of one system’s losing periods are also losing periods for another, treat the pair as overlapping and look for a different leg.

overlap-example

The numbers below are made up to show the method. They are illustrative, not real results, and six weeks is far too short to judge anything. Results are in units, where one unit is one standard stake.

Look at what it shows. In weeks 2 and 5 the lower-priced leg lost while the higher-priced leg won big, which is the kind of cushioning you hope for. In week 3, though, the higher-priced and lay legs lost together and the combined result was the worst of the six. A portfolio does not remove bad weeks. It changes how often they happen and how deep they are, and only your own records can tell you which.

If week 3 kept repeating across a much longer sample, that would be a sign those two legs overlap. Then you would look at what they share, such as the same course type or the same style of race, and change one of them.

Run every system from one shared bank

A portfolio works from one bank. Each leg takes its stake from the same pot, using the same staking rule, so the bank sees the combined result and not three separate stories.

Keep it simple. Choose one staking approach and apply it to every leg. If you need a starting point, a consistent staking plan that keeps each stake a small, fixed share of the bank is easier to live with than anything clever. Then size the whole bank against the worst stretch you could face. The guide to drawdowns and bankroll size walks through how to think about that worst stretch before you commit money.

Two practical rules help:

• Do not let the lay leg run on larger liabilities than your backing legs risk. Cap liability per bet at the same size as the stake you would use on a back.
• Record every bet in one place, tagged with its system name, so you can split the results later. If you cannot separate the legs in your records, you cannot run the overlap check.

Use the bet tracker or a spreadsheet, as long as every bet carries a system tag.

Why a portfolio might survive when a single system would not

Each system has its own worst losing run, and they are not the same size. A higher-priced backing system can go a long way without a winner simply because it bets longer odds. A lower-priced system, winning more often, tends to have shorter dry spells. A lay system, winning most of its bets, tends to string fewer losers together, though each loss costs more.

To see how long runs can get, look at the verified figures above. Over those 13 years, the longest losing run was 28 for the favourite, 36 for the Race Predictor top horse and 59 for VDW Rank 1. None of those means a system is broken. They are what long samples can contain, and on their own, long losing runs are exactly what makes people quit.

The hope with a portfolio is that bad patches do not arrive together, but that has to be shown, not assumed. Systems that share the same race types, market conditions or selection logic can lose at the same time. So place each system’s losing periods side by side and check whether the worst weeks repeatedly land together before you combine stakes. Only treat systems as partners if their bad spells do not keep overlapping.

It also matters beyond the spreadsheet. A lot of betting failure is really the stress of losing runs pushing someone off a sound approach. A portfolio that feels calmer day to day may be one you are more likely to still be running in a year, and staying in the game matters.

Build it one tested system at a time

None of this means launching three untested systems and hoping they balance out. Build the portfolio the way you would build a single system, one step at a time.

1. Get the first system tested live and profitable. Do not add anything until it has behaved as expected over a proper sample.
2. Write its card: the job, the return target, the expected bets per month and the review point.
3. Add the second system and track it separately. Check it earns its place alongside the first, and run the overlap check once you have enough periods.
4. Add the lay leg last, with sensible prices and capped liability. Watch what it does to your overall strike rate and your worst week.
5. Move all three onto one shared bank and one staking rule, with every bet tagged by system.
6. Review on a fixed schedule, not when you feel nervous. The same standard that built your first system that actually works keeps the portfolio honest.

Each piece has to justify itself before the next goes in. Samples under about 1,000 bets are suggestive only, so be patient with each leg.

A word on where systems come from. Many that look brilliant were shaped to fit the past, which is why the backfitting problem deserves a read before you add any new leg. A leg that only works on history is a weak foundation for the whole portfolio.

Know when to leave it alone and when to pause a leg

Once the portfolio is running, the biggest risk is you. The temptation is to fiddle every time a leg has a bad fortnight. Decide the rules while you are calm.

leave-alone

• A leg is inside the losing run or drawdown range you wrote on its card.
• The bad spell is on one leg only and the others are behaving normally.
• You are tempted to change a filter because of the last few results.

pause-or-review

• It breaks the review point you set in advance.
• It starts losing in the same periods as another leg, again and again.
• Its prices or conditions have clearly changed from what you tested, for example you can no longer get the odds the system assumed.

Reviewing is not the same as tweaking. A review asks whether the leg still does the job on its card. Tweaking changes the leg to chase recent results. If you feel the urge to tweak, read why you should stop tweaking your system first, and make any change a proper test, not a reaction.

Common mistakes to avoid

• Adding legs too quickly. Three untested systems is not a portfolio, it is three unknowns.
• Choosing systems that look different but bet on the same horses. Run the overlap check.
• Using different staking rules for each leg, so the bank has no single logic.
• Treating a lay leg as free money because it wins often. The big losses are the whole risk.
• Judging a leg on a handful of weeks. Short samples are noise.
• Changing a leg because of last week. Follow the review rules you set.
• Not tagging bets by system, which makes every later check impossible.

Frequently asked questions

faq-how-many

Two or three. More than that is hard to track and rarely adds protection unless each new system is truly different from the rest.

faq-guarantee

No. It can help only if the systems lose at different times, and that has to be checked in your own records. Bad weeks can still happen together.

faq-lay

No. A two-leg portfolio of a higher-priced and a lower-priced backing system is a fine start. Add a lay leg only when you have a tested one with sensible prices and capped liability.

faq-switch

No. Adding a system to escape a bad patch is a reaction, not a plan. Add a leg only when it has been tested and passed the overlap check.

faq-tracking

Date, race, price taken, stake, result and the name of the system. With those fields you can split the results by leg and run every check in this guide.

Do this today

• Write the job and return target for each system on a short card.
• Decide the losing run or drawdown at which you will review each leg.
• Tag every bet with its system name from now on.
• Pick one staking rule and one shared bank for all legs.
• List each leg’s results week by week and mark the losing weeks.
• Count how often two legs lose in the same week.
• Add a new leg only after the current one is tested.

When you are ready to look at live races, open today’s race cards and sketch which kind of horse each of your three legs would pick.