The Variable Reward Trap: When Bad Trades Payeric

This is part sixteen of the ASX order book series. The apprenticeship curve described phase 3, where traders know what to do and keep not doing it. This post is about one of the main reasons why.

A recurring and very honest theme in the journals I read was some version of breaking every rule on a given day and still finishing green. Sometimes it came with relief, sometimes with genuine worry. The worried ones were right.

Slot machine psychology

Behavioural psychology has a name for the most addictive reward schedule: variable ratio reinforcement. You don't know which pull of the lever will pay, only that some will. It's the schedule slot machines use, and it produces behaviour that is extremely persistent and very hard to extinguish.

Markets run on exactly this schedule. And the specific behaviours they reward unpredictably are the ones that eventually destroy traders:

Each time one of these pays, your brain files it as a success. The lesson learned is precisely the wrong one.

Outcome bias

The cognitive version of the same problem is outcome bias: judging a decision by its result rather than by the quality of the decision at the time. If you run a red light and arrive safely, the decision didn't become good. You were lucky.

In trading, outcome bias is everywhere, because results are noisy. A good trade can lose. A terrible trade can win. Over a handful of trades, there's very little connection between the quality of the decision and the result. Only over many trades does quality show through.

A useful frame is a simple two-by-two:

Most traders only learn from the bottom right box. The traders who improve fastest learn just as much from the bottom left.

What it looked like in the journals

The journals showed the trap in action again and again:

• Best-ever weeks that, on honest review, were built on broken rules and oversized trades that happened to work.
• Stretches of strong P&L while discipline quietly eroded, followed eventually by a loss large enough to erase the stretch.
• Traders openly acknowledging that being rewarded for bad habits made those habits much harder to drop.

The honest ones posted these as warnings to themselves. They knew the bill was coming. It usually did.

How to break the trap

You can't change how markets reward behaviour. You can change what you measure and reward. These are the tools the traders used.

1. Score process separately from P&L

Give every trade, or every day, a process score. Did you trade only your setups? Did you respect your stops and loss limits? Was the size correct for the grade? Did you complete your checklist? Record it before you look at the result if you can.

Over time, track process score and P&L side by side. You want to see that high-process days are profitable on average. If you're making money on low-process days, that's the warning light, not good news.

2. Log rule-breaking trades separately

Tag every trade that broke a rule. Then look at the P&L of that tag alone, over months. In the short run it might be positive. Run a simple simulation and you'll see why that's misleading.

Copy import numpy as np

def rule_breaking_pnl(months=24, trades_per_month=10, seed=3):
    """Rule-breaking trades: usually small wins, occasionally a disaster."""
    rng = np.random.default_rng(seed)
    pnl = np.where(rng.random((months, trades_per_month)) < 0.97, 0.5, -20.0)
    monthly = pnl.sum(axis=1)
    return (monthly > 0).mean(), monthly.sum()

print(rule_breaking_pnl())  # most months positive, total clearly negative

With these made-up numbers, most months look fine. The total is a loss. That's the shape of the trap.

3. Treat lucky wins as errors in your review

When you review your week, put rule-breaking wins in the same section as losses. Write down what rule you broke and what would have happened in the bad case. It feels strange to analyse a winning trade as a mistake. It's exactly the right thing to do.

4. Hide your P&L while trading

One practical trick that came up in more than one journal: hiding the running P&L during the session. Watching the number tick up and down pushes you toward decisions about the number rather than the trade: taking profits too early to lock in a green day, or holding losers to avoid a red one. Without it, you're left with the book and your plan.

5. Make the rule cheaper to follow than to break

Hard stops at the broker. Maximum order sizes. A daily loss limit that locks you out. A daily trade cap. Each of these removes a decision from the moment where the variable reward has the most pull.

The long game

The traders who eventually did well didn't get there by becoming immune to the trap. They got there by recognising it faster each time, and by building systems that made the dumb-luck box smaller. Profit is the scoreboard. Process is the game. Over a long enough run, they converge. Over a short run, they can point in opposite directions, and that's when you find out which one you're really playing for.