Advertisement IC Markets

Variance and the Illusion of Skill

Master exponential money management, position sizing calculators, strict daily stop-loss limits, and overcoming FOMO on micro-timeframes.
Post Reply
Fairman
Posts: 6955
Joined: Tue Jul 21, 2026 7:11 am
Location: Abuja

Variance and the Illusion of Skill

Post by Fairman »

Variance and the Illusion of Skill

Randomness produces patterns that look like skill and patterns that look like failure. Understanding variance protects you from reading too much into results.

Imagine 1,000 traders, each with a strategy that has zero edge, taking a trade a day with a fifty percent chance of winning. After 20 trades, some will have won 14 or more, which is about a 6 percent chance for each trader, so around 60 of them will appear to be successful. After more trades, a few will look like geniuses. They are not. They are lucky outliers in a large population.

The same logic applies to funds, signal providers, and social media profiles. Selection plays a role: you see the survivors, not the many who did equally well by chance and then failed.

In your own results, the lesson is to demand more evidence. A good month is not proof of skill, and a bad month is not proof of failure. Sample size, as discussed, and statistical tests help distinguish.

The practical stance is to treat early success with skepticism, to avoid raising risk after good results, and to avoid abandoning a sound method after a bad stretch.

It also guides how you judge others. A trader with a great record over a short period may simply be lucky. A long, verified record across market conditions is far more informative.

Practical step: simulate a hundred random traders with a coin flip in a spreadsheet, and look at the best and worst results after 50 trades.
It’s Fairman :geek:
Recommended broker for automated trading & scalping IC Markets
Fairman
Posts: 6955
Joined: Tue Jul 21, 2026 7:11 am
Location: Abuja

Re: Variance and the Illusion of Skill

Post by Fairman »

The coin flip spreadsheet is worth doing. Watching a few random traders look brilliant after 50 trades is the fastest way to get suspicious of short track records. In a spreadsheet, put RAND() greater than 0.5 in a grid of 100 columns by 50 rows, give each win plus 1R and each loss minus 1R, and sum each column. Several columns will show impressive totals, and a few will look terrible, all from pure chance. Press recalculate and a different set of stars appears. Once you've seen that, a signal seller showing two good months stops being impressive. You start asking for hundreds of trades and verified records instead.
It’s Fairman :geek:
Fairman
Posts: 6955
Joined: Tue Jul 21, 2026 7:11 am
Location: Abuja

Re: Variance and the Illusion of Skill

Post by Fairman »

Applies to our own best months too. The temptation to raise risk after a hot streak is strongest exactly when luck might be doing most of the work. I'd decide in advance what evidence is needed before risk increases, something like at least 100 trades with positive expectancy and rule following above 90 percent, then increase in small steps. If a great month makes you want to double size, wait another month and see if the feeling survives. Usually the next month is more ordinary, and you're glad you didn't. The same restraint applies to bad months. Don't cut a strategy because of one losing stretch that fits normal variance.
It’s Fairman :geek:
FXS Prop Desk
Posts: 1689
Joined: Wed Oct 07, 2026 9:52 am

Re: Variance and the Illusion of Skill

Post by FXS Prop Desk »

A practical takeaway: when evaluating your own results, compare them with a random baseline. If your win rate over 50 trades is within the range expected from chance for a strategy with no edge, more trades are needed before drawing conclusions.

Keeping track of results over a large sample, and comparing with chance, helps avoid overconfidence after a lucky streak.
FXS Official Team | The Idea Lab & Strategy
Shadow Trader
Posts: 5594
Joined: Wed Sep 09, 2026 5:19 pm

Re: Variance and the Illusion of Skill

Post by Shadow Trader »

Running the same 1,000 zero edge traders out to 100 trades shows how slowly the noise fades. About 184 of them, roughly one in five, finish with 55 or more winners from pure chance.
img_13325_r1.png
img_13325_r1.png (82.05 KiB) Viewed 190 times
The chart (illustration) puts that next to 1,000 traders with a real 55 percent edge at the same 100 trades. The two curves overlap heavily, and about 183 of the genuine traders finish at 50 wins or fewer, which looks like no edge at all.

So after 100 trades, a 55 percent result cannot separate skill from luck. The standard error at that sample is 5 percentage points, and getting the gap to two standard errors needs around 400 trades.

That figure matters for the random baseline idea above. A 55 percent win rate over 50 trades sits inside the zero edge range, so the honest reading is that there is no verdict yet, rather than a sign that the method works.

For a scalper taking ten trades a day, 400 trades is only about two months. Do people find it easier to wait for that sample when they write down the required trade count before the strategy goes live?
Shadow Trader
Posts: 5594
Joined: Wed Sep 09, 2026 5:19 pm

Re: Variance and the Illusion of Skill

Post by Shadow Trader »

Win rate is only half the noise, and the payoff shape decides how much variance sits on top of it. A low hit rate system with large winners is far harder to judge than the coin flip example suggests.

Take a zero edge system that wins 25 percent of the time at 3R against 1R losses. Each trade then swings by 4R between outcomes, and over 100 trades the standard deviation of the total is about 17R, against 10R for a plain 50 percent system at 1R.

So roughly one zero edge trend follower in six finishes 100 trades at plus 17R or better. That record would look convincing to most people, and it came from nothing at all.

The sample needed grows with it. To show an edge of 0.1R per trade at two standard errors, the 50 percent system needs about 400 trades, while the 25 percent version needs around 1,200.

That matters for the signal sellers mentioned above, since many advertise exactly this low hit rate, big winner profile. Does anyone apply a stricter trade count to those records than to high win rate ones?
Shadow Trader
Posts: 5594
Joined: Wed Sep 09, 2026 5:19 pm

Re: Variance and the Illusion of Skill

Post by Shadow Trader »

Imagine a scalper who nets 1.5% a month on a 3,000 dollar account. That's 45 dollars. Spread across 60 hours of screen time and prep, the rate is 75 cents an hour.

Now imagine the same person on a 30,000 dollar account with the same percentage. The monthly figure jumps to 450 dollars, or 7.50 an hour. Still modest, but a different conversation.

The point isn't discouragement. It's clarity. Return percentages that look impressive on small balances translate into small dollar amounts, and growth in capital matters as much as improvements in skill.

That's why many people treat the early years as an apprenticeship. The learning curve is steep, and the income comes later, if it comes at all.

Plans that assume quick wealth from a small account tend to produce risky behavior. Plans that treat it as a long skill-building project tend to survive.

A related exercise: calculate the same figure using only profitable months, then using all months including losers. The honest hourly rate over a full year, with drawdowns included, is usually lower than the highlights suggest. Neither number is a reason to quit, but both are reasons to plan carefully.

It also exposes the value of cost control. Cutting two hours of low-quality trading per day, if those hours broke even, reduces effort without hurting income. Time is the one resource that can't be recovered, and trading consumes more of it than most people track.

Some traders find the exercise liberating. Seeing a poor hourly rate removes the illusion of easy money and shifts attention to building capital and skill, the two things that actually move the number.

What expectations about income had to be corrected, and when?
Post Reply