I’ve been trading for more than six years now, and one of the things that still fascinates me is how difficult it is for us to understand something as simple as probability. Almost everyone who has spent some time in the markets knows that trading is a game of probabilities, that no strategy wins every time and that even a profitable system will occasionally go through losing periods. We hear these things so often that they eventually become almost meaningless, and I think that’s partly because understanding probability on paper is completely different from experiencing it when there’s actual money involved.
I remember when I started trading and genuinely believed that the better I became, the more accurately I would be able to predict what happens next. I thought that after enough years of experience, thousands of hours looking at charts and enough backtesting, I would eventually reach a point where I could look at the market and understand with a high degree of confidence where it was going. Obviously, experience has helped me recognize certain situations and understand what I’m looking for, but the longer I trade, the more comfortable I’ve become with admitting that I have absolutely no idea what will happen next.
And surprisingly, I think that’s one of the things that has helped me the most.
I. Probability is not correlated with previous results
Imagine that you have a strategy with a 35% win rate and an average winner three times larger than your average loser. If those numbers hold over time, before costs, you have a positive mathematical expectancy, even though you will lose approximately two out of every three trades.
Now imagine actually trading that strategy.
You start the week with a loss, take another trade the following day and lose again. Wednesday comes around and you finally get a setup you’ve been waiting for, but that one loses too. By Thursday, you’re already starting to feel slightly uncomfortable and after another loss on Friday, you begin wondering whether something has changed in the market or whether your strategy has somehow stopped working.
Mathematically, four consecutive losses are completely normal for a strategy that loses 65% of the time. Assuming independent trades with the same probabilities, the chance of losing four specified trades in a row is approximately 18%, which isn’t particularly unusual. Yet when those losses happen to you, especially with real money involved, they can feel like evidence that something is wrong.
This is where I think a lot of traders misunderstand probability. They understand that losing streaks are possible, but they don’t really expect to experience them or at least not as frequently as they actually happen. They imagine a profitable strategy producing a relatively smooth equity curve where winners and losers alternate in some reasonable pattern, while the reality can be much messier.
A strategy with a 35% win rate doesn’t owe you a winner after three losses. It doesn’t know how much you’ve lost this week and it certainly doesn’t care that you’re starting to question yourself. If the underlying probabilities haven’t changed, the next trade still has the same estimated chance of winning as before, regardless of whether the previous five trades were winners or losers.
Of course, real markets are more complicated than independent coin flips and probabilities can change with market conditions. But that doesn’t change the underlying point: a losing streak alone doesn’t prove that your strategy has stopped working.
II. I experienced this in a way I wouldn’t recommend
Some of you probably remember when I shared that I went through a streak of 36 consecutive losing trades. I was trading a strategy with a relatively low win rate, so losing streaks were something I had already accepted as part of the process, but there’s a pretty big difference between accepting that something can happen and actually sitting through it for weeks.
I would open the charts, wait for my setup, execute the trade and get stopped out. Then I’d close everything, come back the next day and go through essentially the same process again, only to get stopped out once more. At some point it almost became funny because I genuinely started wondering how many more times I could repeat the same thing before finally getting a winner.
What made the experience particularly interesting was that there wasn’t necessarily some obvious mistake I could point to after every loss. Of course, I reviewed my execution and questioned whether the strategy was still behaving as expected, because blindly repeating something forever without evaluating it would be stupid. But I also understood that changing the strategy simply because I was uncomfortable with the results could easily make things worse.
I think this is one of the most difficult parts of trading a system with a statistical edge. You need enough confidence in your research to continue following the process during difficult periods, but you also need enough humility to accept that your original assumptions might be wrong. You can’t assume every losing streak is just bad luck, but you also can’t treat every losing streak as proof that your strategy is broken.
There isn’t a perfect answer to that problem, which is probably why I find it so interesting.
III. We are terrible at separating decisions from outcomes
Something I’ve noticed over the years is how naturally we judge our decisions based on what happened afterwards. If I take a questionable trade and it makes money, part of my brain immediately wants to believe that the decision was justified. If I follow my plan perfectly and lose, the same brain starts looking for something I should have done differently.
This is often called outcome bias, and it has been studied outside trading as well. In a classic paper, Baron and Hershey (1988) found that people evaluated decisions differently depending on their outcomes, even when the quality of the original decision should have been judged using the information available beforehand.
I think this translates almost perfectly to trading because the market constantly gives us feedback that can be incredibly misleading. You can break every rule you have, take a ridiculous position and make money, which might encourage you to repeat exactly the same behavior tomorrow. Meanwhile, you can execute a statistically sound strategy correctly and lose money for several weeks, which might convince you to abandon something that was working perfectly well over a larger sample.
When I review my own trading, I try to remind myself that I have to judge a decision based on what I knew at the moment I made it, rather than everything that became obvious afterwards. It’s incredibly easy to look at a completed chart and explain what should have happened, but the trader who entered the position didn’t have access to those future candles.
And that’s the person whose decision actually matters.
IV. A small sample can tell you almost anything
I think one of the biggest mistakes traders make is drawing conclusions from ridiculously small samples. Someone backtests a strategy over twenty trades, gets a 60% win rate and starts believing they’ve found something incredible. Then they take the same strategy live, lose five trades in a row and immediately start looking for another system because apparently the edge disappeared.
The problem is that twenty trades tell you very little about the true probability of success, especially when the outcomes are noisy and the strategy depends on changing market conditions.
Let’s take a hypothetical strategy with a true win probability of 40%. If you take twenty independent trades, your observed win rate can easily be quite different from 40% simply because of randomness. You might win eleven trades and think you’re trading a 55% win-rate system, or win only five and conclude that your strategy is terrible, even though neither result necessarily means anything fundamental has changed.
This is also why I don’t particularly care about evaluating my performance based on one week or one month anymore. Of course I track the numbers and pay attention to whether something unusual is happening, but I try not to let a small collection of outcomes determine how I feel about my entire approach to trading.
There is a concept called the law of large numbers, which essentially tells us that under suitable conditions, averages tend to move closer to their expected values as the number of observations increases. It doesn’t mean that every hundred trades will perfectly reflect your theoretical win rate, and it certainly doesn’t guarantee that your historical edge will continue working in the future. But it does explain why judging a probabilistic process based on a handful of outcomes is often misleading.
The difficult part is that we experience trading one trade at a time, while the mathematics only starts becoming meaningful over a much larger collection of decisions.
V. Knowing your probabilities changes how you think about risk
One of the things that changed the most throughout my trading career is how I think about risk. When you’re starting out, it’s incredibly tempting to focus on how much money you could potentially make, especially after finding a strategy that seems to have a statistical edge. You look at the backtest, calculate the expected returns and naturally start thinking about how quickly you could grow your account if you increased your position size. But I think the more important question is whether you could actually survive the losing periods that come with that strategy, both financially and psychologically.
Imagine risking 1% of your account on every trade. That might sound perfectly reasonable, and depending on your strategy and financial situation, it absolutely can be. But ten consecutive losses would leave you with approximately a 9.56% drawdown if you’re risking 1% of your remaining balance each time, and twenty consecutive losses would put you down around 18.2%. Those numbers become much more uncomfortable when you’re actually experiencing them rather than looking at them in a spreadsheet. Suddenly you’re questioning whether the strategy still works, whether you should reduce your risk or whether something about the market has fundamentally changed, even though a losing streak alone doesn’t necessarily tell you any of those things.
I’ve experienced long enough losing periods to understand how differently risk feels when everything is going well compared with when you’ve spent weeks watching your account slowly decline. That’s why I think position sizing should be based on more than just the average return you expect from your strategy. You also need to understand its historical drawdowns, how uncertain those estimates are and whether you could realistically continue following the same process if you experienced something even worse. Because having a profitable strategy on paper doesn’t mean much if the amount you’re risking eventually makes it impossible for you to execute that strategy consistently.
This is why I think understanding probability should directly influence how much you risk. It’s relatively easy to look at a backtest, see the average return and calculate how quickly you could grow an account with more aggressive sizing. It’s much harder to look at the same data, imagine going through the worst losing period while paying your bills and dealing with everything else happening in your life, and then decide whether you could realistically continue executing the strategy without interfering.
I don’t think the best risk level is necessarily the highest amount your account can mathematically tolerate. For me, it also has to be an amount that allows me to remain psychologically consistent when the outcomes aren’t going my way.
VI. The hardest part is accepting that you might never be rewarded
There is one uncomfortable thing about probability that I think deserves much more attention, and that’s the fact that having an edge doesn’t guarantee a particular outcome.
You can have a strategy with positive expectancy, execute it correctly and still finish a month or even a year losing money. The probability of that happening depends on the strategy, the number of trades, the distribution of returns and many other factors, but the possibility never completely disappears.
I think a lot of people subconsciously treat a statistical edge like some kind of promise. If I follow the rules for long enough, the market eventually has to pay me. But that’s not really how probability works, and it’s especially not how financial markets work, because the underlying conditions can change while you’re collecting the sample.
All we can really do is make decisions using the best information available, manage the downside when we’re wrong and keep evaluating whether the assumptions behind our strategy still make sense. Even then, uncertainty remains part of the process.
Personally, I find that much easier to accept now than I did a few years ago. I don’t need to wake up believing that today’s trade will work, and I don’t need to convince myself that the next setup is somehow special because the previous few didn’t go my way. If it meets my criteria, I take it with the risk I’ve already decided on, knowing perfectly well that I might lose again.
I’ve also stopped trying to find an explanation for every single loss. Sometimes there is something useful to learn, sometimes I made a mistake and sometimes the trade simply didn’t work. I don’t think every negative outcome needs to become a lesson, especially when the decision itself was perfectly reasonable.
VII. So how should you actually think about probability?
If there is one thing I’d encourage you to do, it’s to start looking at your trading as a collection of decisions rather than a sequence of individual wins and losses. Before entering, you should have at least a reasonable understanding of what your setup is supposed to achieve, how frequently it has worked historically, how large the winners and losers tend to be and what kind of drawdowns you might experience if the future behaves similarly to the past.
Then you need to accept that none of those numbers tells you what will happen on the next trade. Your 40% win-rate strategy might win the next five trades or lose all five, and neither outcome automatically proves anything about its long-term profitability. What matters is whether you’re following a process with a defensible edge, whether the risk is appropriate and whether the results over time continue to support the assumptions you originally made.
After more than six years of trading, I think understanding probability has made me much less interested in trying to be right about individual market movements. I still make mistakes, I still get frustrated sometimes and I definitely haven’t reached some magical state where losses don’t bother me at all, but I no longer feel like I need every trade to confirm that I know what I’m doing.
Sometimes I’ll do everything correctly and lose, and sometimes I’ll make a mistake and get rewarded for it. Both are going to happen again, probably many times, and neither one should have enough influence to completely change the way I approach the next opportunity.
I think learning to live with that uncertainty is one of the most useful things trading has taught me, and ironically, it has also made the whole process much less stressful.
Bless you all.
Luke