Ten Losses in a Row Is Not a Signal: Variance for Normal People

A bettor has done most things right. They found a strategy with a sensible idea behind it, checked the sample size, compared the numbers against the league baseline and the odds, and started placing small, flat stakes. For the first few weeks it works. Then comes a Sunday evening with the seventh losing bet in a row, and a quiet, reasonable-sounding thought: it’s stopped working.

So they stop. Or they change the filters. Or they double the stake to “get it back.” Each of these feels like a decision. Each of them is a reaction to a run of results that, statistically, tells them almost nothing.

Seven losses in a row did not mean the strategy was broken. It meant the strategy was doing exactly what strategies with a real edge do — including, regularly, the part nobody puts in the advert.

This article is about that timescale. How long “normal bad” actually lasts, why the human brain reads it as a signal, and three small habits that stop a perfectly good approach from being abandoned at the worst possible moment.

1Meet a Perfectly Good Strategy

To keep things concrete, imagine a strategy with the following profile. It wins 45% of its bets, at average odds of 2.30, with a flat stake of one unit per bet.

The arithmetic is simple. Out of every 100 bets, it wins 45 and collects 1.30 units profit on each — 58.5 units. It loses 55 and gives up one unit on each — 55 units. The difference is +3.5 units per 100 bets, a yield of +3.5%.

This is deliberately a good strategy — better than most people will ever find, and far more honest than the yields advertised on social media. If even this one produces painful stretches, then every strategy does.

2What That Strategy Actually Does to You

Rather than argue about it, we simulated it. The strategy above was run 20,000 times over 100, 500 and 1,000 bets — the same edge every time, only the luck changing. Here is what a bettor using it should expect to live through:

Read the first line again. A strategy with a real, positive edge finishes its first hundred bets in the red more than a third of the time. Not because anything went wrong. Because a hundred bets is a small sample, and small samples, as the stats-table article put it, simply have nothing to say.

The losing runs are the part that does the damage. Over just 100 bets, there is roughly an even chance that this strategy produces a run of seven or more losses in a row. Over a thousand bets, a run of ten is more likely than not. These are not disasters. They are the weather.

The drawdown figures matter just as much. “Drawdown” simply means the distance from your best point to your worst point afterwards. A typical worst drawdown of 22 units over 500 bets means that somewhere in those bets, a bettor with a genuine edge will watch 22 stakes disappear before the line recovers. If each stake was £20, that is £440 of “proof” that the strategy is broken — from a strategy that isn’t.

3Why Your Brain Reads Noise as News

None of this is intuitive, and there is a good reason for that. Human beings are pattern-detectors first and statisticians a very distant second. Five losses in a row feels like information: something has changed, the market has adapted, the league has shifted, the method has been “found out.”

Sometimes one of those things is true. Most of the time, the run is just what a 55% loss rate produces when it is allowed to happen in public, one bet at a time.

The same mistake appears in a different costume on stats tables, where a team’s 5W streak looks like momentum. A streak tells you what happened. It does not tell you why. Your own betting record is no different: a run of results — good or bad — is a prompt to look closer, never a verdict on its own.

There is also an asymmetry worth naming. Wins are remembered as skill; losses are remembered as events. Three wins in a row feel like confirmation. Three losses feel like a warning. Neither deserves either label.

4The Opposite Trap: The Lucky Start

Variance cuts both ways, and the second direction is more dangerous because it feels so good.

Take a strategy with no edge — worse, a strategy that loses 5% per bet over the long run (a 41.3% strike rate at the same 2.30 odds). Run it through the same simulation. After 100 bets, roughly one bettor in three is in profit. After 500 bets, still roughly one in six.

This is the quiet engine behind a great deal of disappointment in betting. Someone finds a set of filters, enjoys a good first month, increases the stake because “it’s clearly working,” and meets the long run at the larger stake. The early profit was never evidence. It was a coin that happened to land well while they were watching.

The honest conclusion is uncomfortable but useful: over a few dozen or even a few hundred bets, your results alone cannot tell you whether you have an edge. Good and bad strategies produce overlapping outcomes for a long time. That is precisely why the question has to be answered with something other than the bankroll balance.

5Three Habits Instead of Panic

None of the above requires a statistics degree to act on. It requires deciding a few things in advance, while you are calm, rather than in the middle of a losing run, when you are not.

Decide what “normal bad” looks like before you start

Before the first bet, write down two numbers: the longest losing run you should expect, and the worst drawdown you should expect, for the number of bets you plan to place. For the strategy above over 500 bets, that is roughly nine losses in a row and twenty-odd units down from the peak. If those things happen, nothing has happened. You were told in advance.

Judge after a fixed number of bets, not a feeling

Choose the review point before you begin — 200 bets, 300, a full season — and do not re-judge the strategy before you reach it unless something genuinely changes outside the results (the league restructures, the market you bet into disappears, you realise you have been breaking your own rules). A losing week is not a review point. Neither is a winning one.

Size the stake for the bad case, not the average one

If the one-in-ten drawdown for your plan is around 50 units, ask a simple question: would losing 50 stakes, temporarily, still be fine — financially and emotionally? If not, the stake is too large, regardless of how good the strategy is. A strategy you abandon at the bottom of a drawdown has the same practical value as a strategy that never had an edge.

6What Software Can Do, and What It Can’t

A good analytical tool can make most of this visible before it happens, which is the whole point. In CGMBet26, the Staking tab of the Backtest window takes the strategy’s own bets, runs them through thousands of simulated histories, and shows the spread of outcomes — the longest losing run, the typical and the bad drawdown, the chance of being in profit at all — not just the average line. The Bets Tracker keeps an honest count of where you actually are, so the review point arrives with real numbers rather than impressions. And the Robustness tab asks the harder question underneath all of it: whether the edge in the backtest was ever likely to be real, or whether the filters simply found a lucky stretch.

What no tool can do is sit through the losing run for you. The simulation can tell you, in advance and in plain numbers, that nine losses in a row is ordinary. It cannot make the ninth one feel ordinary on a Sunday evening.

7The Shortest Version of All of This

Every strategy, good or bad, produces long stretches that look like the opposite of what it is. Good strategies lose for weeks. Bad strategies win for months. Results over a small number of bets are not evidence of either, and the most common way bettors lose a genuine edge is not by never finding one — it is by abandoning it in the middle of a perfectly normal losing run.

Decide your “normal bad” in advance. Judge at a fixed point, not a feeling. Keep the stake small enough that the bad case is dull. Everything else is commentary.