It's 2:47 AM. Your NinjaTrader bot entered a long position 14 minutes ago. It's currently −$380 against you. The backtest says the average trade has a drawdown of −$210 before reaching its profit target. This trade is running hotter than average. Your hand is on the mouse.
This is the moment that separates successful algorithmic traders from unsuccessful ones. Not the code. Not the backtest. This moment.
I've been here dozens of times. I've closed positions early and watched them hit the profit target. I've let the bot run and watched it recover exactly as the backtest suggested it would. The data on what to do is unambiguous. The psychology of doing it is not.
Why we override our own bots
Understanding the psychological mechanism makes it easier to design systems that work around it. There are three primary override triggers I've identified in myself and in other algo traders I've spoken with:
Loss aversion in real-time
The backtest ran 312 trades over 3 years. When you're watching trade 47 in real-time, your brain doesn't have access to the 311 other trades. It has access to this trade, right now, and the −$380 that exists in this moment as a real number representing real money. The backtest's expectancy is an abstraction. The loss is concrete.
Behavioral economists call this myopic loss aversion: when we evaluate outcomes too frequently, short-term losses loom larger than they should relative to the long-run expectancy of the strategy. The cure isn't discipline through willpower. It's systems that reduce evaluation frequency.
The illusion of control
Trading attracts people who like to be in control. The appeal of building an algo is, paradoxically, that you remain in control through the design phase. But the moment you let it run, you surrender control to the system. For many traders, that's unbearable — especially during a drawdown, when overriding the bot feels like "doing something" rather than passively accepting a loss.
The override is rarely about profit optimization. It's about regaining a sense of agency. Recognizing that is the first step to managing it.
Post-hoc rationalization
The most dangerous override isn't the emotional one — it's the rational-sounding one. "I'm going to close this trade because I see a news event I didn't account for." "The market structure has changed since I built this strategy." These stories are sometimes true. They're more often our brains constructing justifications for what fear wanted to do anyway.
Every time I've overridden my bot with a compelling-sounding reason, I've been right about 30% of the time. The bot, left alone, is right 54% of the time. The math is not subtle.
The five rules I use
Never watch a live trade in real-time
Set the position. Close the chart. If you can't see the unrealized P&L, you can't override it. NinjaTrader lets you minimize or close individual chart windows while the strategy continues to run. Use this feature aggressively.
Write down your override criteria before the session starts
There are legitimate reasons to override: a true system error, a broker outage, a verified news event that your strategy explicitly excludes. Write these down before the session. If your reason for overriding isn't on the pre-written list, it doesn't count.
Track every override as a separate P&L line
Every time you override the bot, log it. Record what the bot would have done vs. what you did. After 20 overrides, look at the data. This is usually sufficient to cure the habit permanently — the numbers are humbling.
Separate "strategy maintenance" from "live session" mentally
Legitimate improvements to a strategy happen in the lab: new backtests, parameter updates, code changes. None of those happen during a live session. If you find yourself wanting to "adjust" something while a trade is running, write it down and address it after the session closes.
Use hard circuit breakers, not soft intentions
The only reliable substitute for willpower is removing optionality. Set your daily max-loss in NinjaTrader's risk settings so the platform closes your positions for you if things go catastrophically wrong. This way, the only legitimate reason to override manually is a system failure — not emotion.
The night-shift advantage nobody talks about
Here's the counterintuitive truth about trading while you sleep: it's easier to maintain discipline at 2 AM than at 10 AM. When you're fully awake, alert, and at your desk, your brain is maximally engaged — and maximally vulnerable to the override impulse. When your strategy is running while you sleep, the decision to not interfere has already been made. You don't override what you don't see.
This is part of why I built my primary strategies around the overnight session. Not just because of the edge in the data — but because removing myself from the equation during live trading is genuinely easier when I'm asleep. The bot becomes less of a tool and more of a trusted partner, operating in a window I've explicitly handed to it.
The best trading journal entry I ever wrote was: "I slept through the entire session. Bot took 2 trades. +$840. I had nothing to do with it."
Scale your discipline with funded capital
One more reason to use prop firm funded accounts: the evaluation rules actually help with discipline. When you know breaking the daily loss limit ends your funded account, you're far less likely to override your bot to "just see what happens." The external constraint reinforces the internal one.
Bulenox
90% profit split, no consistency rules, NinjaTrader automated trading fully supported.
Start Evaluation →TradeDay
Day-one payouts, no consistency rules, NinjaTrader compatible. Up to 6 funded accounts.
Get Funded →Disclosure: Bulenox and TradeDay links are affiliate links. Commission earned at no cost to you. Trading futures involves substantial risk of loss.