Loss Aversion in Trading: Why You Cut Winners and Hold Losers

You move the stop on a losing trade without thinking. But taking profit early on a winner feels like "being sensible".

This is a plain-English guide to loss aversion in trading.

Same brain. Same account. Completely different standards.

That, in one sentence, is loss aversion in trading.

Why a £100 loss hurts more than a £100 win

Behavioural finance has an awkward finding: people hate losses more than they like gains, roughly by a factor of two.

Gain £100, you feel good. Lose £100, you feel disproportionately awful.

Daniel Kahneman and Amos Tversky modelled this in prospect theory. Simplified, the value function for outcomes looks something like this:

The exact multiplier varies across studies, but the direction is consistent: losses loom larger than gains.

Your brain isn't measuring P&L. It's measuring discomfort.

Put that brain in front of a trading screen, and you get perfectly irrational behaviour.

How loss aversion in trading flips your edge upside down

Imagine a simple hypothetical system.

It wins 50% of trades. Average win +2R. Average loss −1R. No fancy tricks.

On paper, expectancy per trade is:

E = (0.5 × 2R) + (0.5 × −1R) = 1R − 0.5R = +0.5R

Risk 1% per trade and the average outcome is +0.5% per trade.

Nothing extreme. Just a clean, asymmetrical payoff.

Now overlay typical loss-averse behaviour on top of that system:

• You cut winners early at +0.8R because "it might reverse"
• You move stops wider or remove them once you're at −1R because "the level will hold"
• You occasionally average down to "get a better price"

Let’s be conservative and say your behaviour alters the averages to this:

• Average win drops from +2R to +1R
• Average loss increases from −1R to −1.5R

Win rate still 50%. Trades still entered at the same places. But the new expectancy is:

E = (0.5 × 1R) + (0.5 × −1.5R) = 0.5R − 0.75R = −0.25R

Exact same entries. Now a losing system.

Markets didn't change. You did.

Why cutting winners feels smart and cutting losers feels stupid

If you zoom into the moment, your behaviour makes emotional sense.

You're up +1R on a trade. Closing it gives you a certain gain. Holding it risks turning a sure win into a loss.

Your brain runs a silent calculation: "I can lock this in and feel good, or risk pain. I'll take the biscuit now, thanks."

When you're at −1R, it's flipped.

Closing now “locks in a loss”. Horrible.

Holding keeps alive a small hope that you get back to breakeven and escape the emotional hit entirely.

So you grab small wins and babysit losers.

Over hundreds of trades, that's exactly backwards.

Systematic trading lives or dies on the distribution of outcomes, not the comfort of any single one.

Expectancy doesn’t care how you feel today

There are only three levers that set your long-term trading result:

• Win rate (percentage of winning trades)
• Average win size
• Average loss size

We covered this trade-off properly here, but the summary is simple.

To have a positive expectancy, the combination of those three has to be favourable. That's it.

Loss aversion in trading quietly pushes two of them the wrong way: it shrinks average wins and inflates average losses.

And then traders stare at the equity curve and say "market makers are hunting my stops".

No. Your own psychology is hunting your stops.

The sunk cost trap: why you babysit losers

Once you've sat with a losing position for a while, the pain isn't just the loss.

It's the time. The attention. The extra charts you've read convincing yourself "the analysis is still valid".

Closing it means admitting all of that was for nothing.

That’s the sunk cost fallacy on top of loss aversion.

The more you've invested in a bad idea, the harder it is to dump it.

This is why averaging down feels emotionally easier than taking the hit and walking away.

Mathematically, it's nonsense.

If the expected value of continuing the trade is negative, putting more money into it is just buying more of a bad bet.

But the brain doesn't evaluate expected value; it just wants out of pain.

What loss aversion in trading does to your equity curve

Look at a typical discretionary equity curve wrecked by loss aversion.

• Lots of tiny, controlled wins
• Frequent small-ish losses
• Occasional large, account-bending losers

The ratio of largest loss to average loss gets silly.

A system that "risks 1%" per trade on paper suddenly has −6%, −10%, even −20% events when fear and hope override the stop.

You don't need many of those to turn a survivable drawdown into a terminal one.

If you haven't read it yet, park a tab for this piece on risk per trade.

Because loss aversion plus oversized risk is how apparently "safe" systems disappear in a month.

Why you can't simply "try harder" to be rational

Most traders respond to this by promising themselves they'll be more disciplined.

They won't.

Not because they're weak, but because the hardware is wrong for the job.

Your brain evolved to survive immediate physical threats, not to optimise long-run expectancy across 1,000 trades.

Fear spikes when you see open losses, not when you see a subtle drift in your risk-reward profile.

Hope spikes when you see green P&L you could "lock in".

You can white-knuckle it for a while.

Then you get the inevitable losing streak, and the animal brain grabs the wheel again.

We covered this loop in more detail in why traders abandon winning systems at the worst time.

The answer is not to become a different person.

The answer is to stop putting your worst instincts in charge of live orders.

What automation fixes (and what it absolutely doesn’t)

Automation gets romanticised as a magic solution.

It isn't.

An automated system will execute a bad edge with perfect efficiency.

What automation can do is brutally simple.

• It follows the entry rules you already defined
• It follows the exit rules you already defined
• It sizes positions exactly as specified
• It ignores how you feel today

In the context of loss aversion in trading, that last line is the point.

Automated execution doesn't care that this is your third loser in a row.

It doesn't tighten the take profit on trade four because "you need a win".

Equally, it doesn't move the stop because gold had a "fakeout".

If your rule is −1R, it dies at −1R.

Cold. Boring. Correct.

What automation does not do is fix a system with no edge.

If your stop is miles away and your take-profit is tiny, a robot will happily grind your account down without ever once feeling guilty.

So you still need the maths on your side first.

Gold, small accounts, and why it feels worse than it is

Loss aversion gets louder when the swings look big relative to your account.

Gold (XAUUSD) is a good example.

On a small balance, a 0.01 lot minimum means each dollar move in gold is a noticeable blip on your equity.

This distorts your perception.

A perfectly normal system fluctuation feels like an emergency because the percentage swings are visible, even when the underlying risk model is sensible.

So you interfere more, cut more winners, and sit on more losers.

This is why system testing, Monte Carlo work, and looking at typical drawdowns before you go live matters.

If you can see, statistically, that "this system routinely takes ten losers in a row and still works", you’re less likely to panic on loss number four.

If you haven't met Monte Carlo before, start here.

Risk-first rules that blunt loss aversion

You can't delete loss aversion.

You can box it in.

That means rules that are boring on a good day and life-saving on a bad one.

1. Predefine exits on both sides

Before entry, you should know:

• Where the stop is (price and R)
• Where initial profit targets are, if you use them
• What conditions, if any, justify a discretionary override

Then the hard bit: you don’t move the line because of how you feel.

If you want to trail stops or scale out, that’s fine.

Just turn it into an explicit, testable rule instead of a vibe.

2. Separate strategy logic from execution

Do your thinking when the market is closed.

Write the rules. Test them. Abuse them with variations.

Then, when the market is open, your job is to follow what you've already decided.

Automation is simply the extreme version of this.

Strategy logic in one box. Execution in another.

The less manual clicking between the two, the less room loss aversion has to improvise.

3. Position sizing that assumes you will have losing streaks

You will have losing streaks.

Loss aversion makes each hit feel worse than the last.

If you size as if you'll never see ten losers in a row, you'll panic when you see five.

A sensible starting point is to risk a small percentage of equity per trade and model what happens over a few hundred trades at that risk.

There's a full breakdown of sizing choices in fixed fractional vs fixed lot, but the guiding question is simple: "Can I emotionally and financially survive the worst plausible run?"

If the honest answer is no, your loss aversion will drive the bus the first time the equity curve dips.

Why your journal is full of hindsight bravery

Open any trading journal and you'll see the pattern.

• Notes like "should have held that winner longer"
• Or "moved stop, took bigger loss than planned"
• Or "exited loser, then it reversed"

On paper, the solutions are clear.

In real time, the same trader repeats the same mistakes because the emotional payoff is immediate and the statistical damage is invisible.

The human brain is great at rewriting history to make last week's panic look reasonable.

This is why systematising helps even discretionary traders.

You don't need to automate everything to get benefit.

Even a semi-automated exit manager that handles stops and targets according to rules can remove 80% of the sabotage.

Loss aversion isn’t a bug. Treat it like volatility

You wouldn't stare at a chart and demand that price stops fluctuating.

You accept volatility and design a system around it.

Loss aversion in trading is the same: it's a structural feature of human psychology.

The job isn't to become emotionless.

The job is to make sure those emotions are not in charge of trade-level decisions.

Understand the edge, understand the risk, then decide what the machine will do and what, if anything, you are allowed to touch.

And if you're consistently cutting winners and holding losers, don’t invent a market conspiracy.

Your behaviour is explainable, predictable, and fixable.

Just not by promising yourself you’ll “be stronger” on Monday.

Trading involves real risk. You can lose part or all of your capital, and no approach removes that risk.

Start your free 14-day ArcisTrade demo →

Watch rule-based systems run, see every winner and every loser, and decide calmly how you want your risk managed before you go live.