That difference matters more than it sounds. Decades of research in behavioural finance show that people make predictable mistakes with money: they sell winners too early, hold losers too long, trade too often and chase losses. Rules written in advance cannot remove the market's risk, but they can remove many of these mistakes. They bring problems of their own: rules can be fitted to the past, can miss news they were never designed for, and can fail quietly when the market changes. This guide compares the two approaches, explains the biases that rules help with, sets out where systematic trading falls short, and shows how many traders combine the two.
The two approaches
| Discretionary trading | Systematic trading | |
|---|---|---|
| Decisions | Made trade by trade | Made in advance as rules |
| Execution | Usually manual | Often automated |
| Testing | Hard to test honestly | Can be backtested and forward-tested |
| Consistency | Varies with mood and fatigue | The same every time |
| Adapting to new information | Fast and flexible | Only as the rules allow |
| Record keeping | Depends on the trader's notes | Every decision is logged |
| Main risk | Emotion, inconsistency | Overfitting, rigid rules, technology failures |
A short history of the debate
For most of the history of markets, trading was discretionary: a trader read the tape, the news and the mood of the floor, and decided. Rules existed as rules of thumb, passed from one trader to another.
Systematic trading began with trend followers. In the 1950s and 1960s Richard Donchian wrote down exact rules for commodity futures, such as buying when the price made a new high over a set number of weeks, and followed them without second-guessing. In 1983 the commodity traders Richard Dennis and William Eckhardt settled a bet about whether trading could be taught: they recruited a group of novices, taught them a written breakout system, and gave them real money to trade it. The "Turtles", as they came to be called, showed that a written rule could be followed by people with no trading background, and also that many of them struggled to follow it when it was losing.1
Meanwhile psychologists were explaining why people find it hard to follow rules. In 1979 Daniel Kahneman and Amos Tversky published prospect theory, which showed that people feel losses more strongly than equal gains and take more risk to avoid a loss.2 In 1985 Hersh Shefrin and Meir Statman named the disposition effect, the habit of selling winners too early and riding losers too long.3 Terrance Odean found it in the records of thousands of US brokerage accounts in 1998,4 and Brad Barber and Odean showed in 2000 that the households that traded most earned the least, largely because of costs.5
Computers turned rules from a discipline into a machine. From the 1990s, funds ran their rules on computers that could watch many markets at once, and by the 2010s retail traders in India could do the same through broker APIs and algo platforms. SEBI's framework for retail algo trading, in force for every broker since 1 April 2026, now sets how such rules reach the exchange.6
Behavioural biases that cost traders money
Research in behavioural finance has documented patterns that cost investors money. Rules help with each of them.
The disposition effect
Investors tend to sell winners too early and hold losers too long. A loss that is not yet booked feels like it might still come back; a gain feels safer once it is taken. The result is a portfolio of small gains and large losses. A stop loss and a target written in advance remove the decision from the moment when the pull is strongest.
Overconfidence and overtrading
Traders tend to overrate their own skill, especially after a run of wins. Barber and Odean's study found that the households that traded most earned the lowest returns, largely because each trade cost money. Rules that require a defined signal for every trade and limit the number of trades a day curb this.
Loss aversion and revenge trading
Losses feel worse than equal gains feel good. After a loss, traders often take larger or riskier trades to win it back the same day. A daily loss limit enforced by software stops this: once the day's loss reaches the limit, new orders are refused, whatever the trader feels.
Hindsight and recency
Recent events feel more likely to repeat than they are, and past events look more predictable than they were. A trader who just lost on an expiry-day spike may avoid expiry days for months; one who caught a big move may expect the next one tomorrow. Testing rules over many years puts recent experience in context.
Anchoring
Traders anchor on a price they have seen: the price they bought at, last week's high, a round number. Decisions then depend on that anchor rather than on the current situation. Rules based on measured conditions, such as an indicator level or a percentage move, avoid the anchor.
What the numbers say for Indian traders
SEBI's studies of individual traders in equity futures and options have found that most lose money. Its study of financial year 2021-22 found that 89% of individual F&O traders made a loss,7 and its study covering 2021-22 to 2023-24 found that 93% of more than one crore individual traders made losses over the three years.8 These studies do not split traders into systematic and discretionary, and a rule-based approach does not make a losing strategy win. What they do show is that the average trader's results are poor, and the biases above are among the reasons.
What systematic trading does well
- Consistency. The same conditions produce the same action, on a good day and on a bad one.
- Testability. Rules can be backtested on past data, forward tested on live prices and compared with each other. See how to backtest trading strategies.
- Speed. Automated execution acts the moment conditions are met, which matters for multi-leg option entries and for fast exits.
- Scale. One person can run several strategies across instruments without watching each one.
- Records. Every entry, exit and refusal is logged, which makes review honest. A trader cannot remember a strategy as better than it was.
- Risk limits that hold. Stops, position sizes and a daily loss limit are applied every time.
Where systematic trading falls short
- Overfitting. Rules can be tuned until they fit the past perfectly and then fail on new data. See avoiding overfitting.
- Rigidity. Rules cannot read news they were not designed for. A sudden policy change, a market halt or a company event may need human judgement.
- Regime change. A strategy that suited one kind of market, such as a calm range, can stop working when the market changes, such as a period of sharp trends.
- Technology risk. Bugs, broker outages and expired broker sessions can leave positions unmanaged. Good systems refuse orders rather than guess when something is wrong, but the trader still has to check.
- False confidence. A tested rule can feel safer than it is. A backtest is one path through the past, not a guarantee.
- Rule changes in the market. Lot sizes, expiry days, trading hours and taxes have all changed in India since 2024. A rule built on the old market may not suit the new one.
What discretionary trading does well
- Context. An experienced trader can weigh news, market mood and unusual conditions that no rule anticipated.
- Flexibility. A trader can change approach the same day the market changes.
- Pattern recognition in situations too complex, too rare or too new to write as code.
- Restraint. A trader can choose not to trade at all on a day that feels wrong, which a rule will do only if it was written to.
Its weakness is that judgement is hard to verify. A discretionary trader cannot honestly test what they "would have done" in the past, because they already know what happened. And their results often depend on how they feel on a given day: tired, angry after a loss, or overconfident after a win.
The hybrid approach
Many professional traders combine both:
- Discretion chooses the strategy; rules execute it. The trader decides which strategies suit current conditions; the algo runs each one exactly as written.
- Rules handle risk; discretion handles ideas. Stops, position sizes and daily limits are automatic, whatever the trader thinks in the moment.
- Discretion can switch rules off, not override them mid-trade. A trader may pause a strategy before a major event, such as a policy announcement or a results day, but does not move a stop during a trade because it feels close.
- Every manual decision is written down, with the reason, so it can be reviewed like a rule.
This keeps the benefits of judgement while removing the most costly emotional decisions.
A worked example: the same idea, two ways
Take a trader who likes to sell a Nifty straddle on expiry morning.
The discretionary version. Each Tuesday the trader looks at the market at 9:30 am. If it feels calm, they sell the at-the-money call and put. They watch the position and exit when it "looks right" or when they get nervous. Over a month, they take four trades: one exited early at a small profit because the market twitched, one held through a sharp move because they hoped it would reverse, two closed well. The month's result depends heavily on the one trade they held too long, and they cannot say which part of their decision-making helped and which hurt.
The systematic version. The same trader writes the idea as rules: sell the at-the-money call and put at 9:20 am on expiry day, stop each leg at 30% above its entry price, exit everything at 3:10 pm, and never trade if the day's loss limit is already reached. They backtest the rules over two years, forward test them for a month, and then run them live at one lot. Every trade follows the same rules. When a month goes badly, the log shows exactly which trades lost and why, and the trader can test a change on past data before making it.
The systematic version is not guaranteed to make more money. What it gives is a strategy that can be measured, improved and trusted to behave the same way next week.
A worked example: rules that protect a discretionary trader
A trader picks stocks by judgement and does not want to automate entries. They can still put rules around the decisions that hurt most:
| Problem | Rule | How it is enforced |
|---|---|---|
| Holding losers too long | Every position has a stop when it is opened | A GTT or a stop order placed with the entry |
| Chasing losses after a bad morning | No new orders after the day's loss reaches ₹10,000 | A daily loss limit on the account |
| Overtrading | No more than 6 orders a day | A cap on orders a day |
| Big moves against the whole account | Close everything at a total loss of ₹25,000 | An account-wide stop that squares off |
None of these rules picks a trade. All of them remove a decision from the moment when emotion is strongest.
Where the biases show up in options trading
Options make the biases above more expensive, because positions move fast and time works against some of them.
- Holding a sold option that has gone wrong. A seller who sold a call for ₹100 and now sees it at ₹250 often waits for it to come back, because booking the loss feels final. The option can keep rising until expiry. A stop on each sold leg, decided before the trade, is the rule that prevents this.
- Adjusting instead of exiting. Rolling a losing leg further away, again and again, can feel like managing risk while in fact adding to it. A rule that limits the number of adjustments, or exits once a loss level is reached, keeps adjustment from becoming denial.
- Buying cheap options for a big win. Far out-of-the-money options are cheap and occasionally multiply. The few wins are remembered; the many small losses are not. A written record of every such trade shows the true result.
- Trading every expiry because of one good day. A trader who caught a large expiry-day move may trade every expiry afterwards, expecting another. Testing expiry-day rules over a year shows how often the big day really comes.
- Doubling size after losses. Doubling up to recover a loss turns a bad day into a bad month. A fixed size, or a size tied to a fixed share of capital, removes the temptation.
Reviewing each approach honestly
Both approaches need review, but they need different reviews.
Reviewing a systematic strategy
| Measure | What it tells you |
|---|---|
| Live results against the backtest for the same period | Whether the rules behave on live prices as they did on past data |
| Slippage per trade | Whether fills are worse than the test assumed |
| Number of trades | Whether the rule is trading as often as expected; too few or too many points to a fault |
| Largest loss and longest losing run | Whether the risk is within what you planned for |
| Refused or failed orders | Whether broker sessions, margin or limits are getting in the way |
Reviewing discretionary trading
| Measure | What it tells you |
|---|---|
| Results by setup | Which of your trade ideas earn and which do not |
| Results by time of day and day of week | Whether you trade worse when tired or on certain days |
| Average win against average loss | Whether you cut winners short or let losers run |
| Trades taken against your own plan | How often you broke your own rules, and what it cost |
| Results after a loss | Whether you trade worse after losing |
A discretionary trader who keeps this record honestly for three months usually finds one or two setups that work and several that do not. The ones that work are the first candidates to write down as rules.
Building rules from your own trading journal
A trading journal is the bridge between the two approaches. A simple method:
- Record every trade for at least 50 trades: the instrument, the time, why you entered, where your stop was, why you exited, and the result after costs.
- Group the trades by reason. For example: "breakout above the morning high", "pullback to the 20-period average", "expiry-day straddle".
- Compare the groups. One or two groups usually carry most of the profit, and some groups lose steadily.
- Write the best group as a rule. "Buy when the price closes above the first 30 minutes' high on 5-minute candles, stop below the first 30 minutes' low, exit at 3:00 pm."
- Test the rule on past data you have not traded, to see whether it holds outside your own trades.
- Drop the losing groups, or keep them only as paper ideas until they show a reason to return.
The rule that comes out of this is yours: it is built from what you already do well, not borrowed from someone else.
Signs that a rule has stopped working
No rule works for ever. These signs suggest a systematic strategy needs review, and possibly a pause:
| Sign | What it may mean |
|---|---|
| A drawdown deeper than any in the backtest | The market has changed, or the test was too optimistic |
| Far fewer or far more trades than usual | The conditions the rule needs have changed |
| Live results much worse than the backtest over the same weeks | Fills, slippage or rule differences on live prices |
| Wins getting smaller while losses stay the same | The edge is fading, often because others trade the same idea |
| A rule change in the market, such as a new expiry day or lot size | The rule was built for a market that no longer exists |
The answer is rarely to change the rule straight away. Pause it, test the change on data, and run the changed rule in a forward test before it trades with money again. That review is itself a rule: it stops a losing week from turning into a rushed redesign.
Moving from discretionary to systematic
If you trade by judgement and want more structure:
- Write down what you already do. Turn your best setups into exact rules: the instrument, the condition for entry, the stop, the target and the time.
- Backtest them. See which of your instincts hold up over many trades and which only felt right.
- Automate the risk rules first: stops, sizes and a daily loss limit, before any entries.
- Forward test the full rule set on live prices with test money, then compare it with your manual results for the same weeks.
- Automate the strategies that test well, starting with the smallest size, and keep reviewing.
The trading automation checklist lists what to check before an automated strategy goes live.
Questions to ask before you automate an idea
Not every idea should become an algo. Before writing one, answer these:
- Can I state the entry in one sentence that a stranger could follow? "Buy when it looks strong" cannot be automated; "buy when the 15-minute close is above the day's opening range high" can.
- Do I know where I am wrong? Every rule needs a stop or an exit condition that is decided before the entry.
- Does it depend on information the rule cannot see? An idea that needs you to read a company announcement or a policy statement may be better kept discretionary, with automated risk limits around it.
- How often does it trade? A rule that trades twice a year cannot be tested well. One that trades many times a day is very sensitive to costs and slippage.
- What does it cost per trade? Brokerage, STT and other charges are the same for a rule and for a person, but a rule may trade more often.
- What happens if the broker session expires or an order is refused? A good system refuses to guess and tells you; you need to know how you will respond.
- Would I keep running it through its worst month in the test? If not, the size is too large or the rule is not one you trust.
An idea that passes these questions is a good candidate. One that fails several is better kept as a discretionary trade with hard limits around it, or dropped.
Doing it in QuantPower
QuantPower is software for traders who want to make their trading systematic and rule-based. It does not tell you which strategy to trade; it helps you write your own rules down, test them and run them.
- Write rules without code. The strategy builder and the algo builder turn legs, entry times, stops, targets and exit times into an algo. Ready-made templates such as straddles and iron condors load as open samples that you change to your own rules.
- Or say them in words. Quanty, the AI assistant in the app, builds a strategy from your description for you to check, and saves it switched off. Only you switch it on.
- Test before money. Backtests run your rules on past data; Forward test runs the same algo on live prices with test money.
- Risk rules above everything. Smart Exit squares off every position and stops the algos at a set loss or profit for the whole account; the daily loss limit refuses new orders once the day's loss reaches it; the kill switch stops trading in a segment until its next session; and a cap on orders a day limits overtrading. These apply to manual orders too, so a discretionary trader can use them on their own.
- A full log. Every algo keeps an activity log of each entry, exit and refusal, and has its own live performance report beside its backtest.
A live algo costs 200 tokens per 28 days, a Forward test 100, and a backtest 2 tokens after the first 10 free each week (1 token = ₹1, GST extra). The risk controls cost nothing.
Common mistakes
- Automating a strategy you never tested. Automation runs a bad rule faster and more often.
- Overriding the rules mid-trade. A trader who moves stops "just this once" is back to discretionary trading, without the benefits of either.
- Switching strategies after every loss. Every strategy has losing runs. Judge a rule on a large sample, not the last week.
- Testing on the same data you designed on. The rule will look better than it is.
- No risk rules for manual trading. Discretionary traders often need hard limits most, and use them least.
- Treating the log as optional. Without a record, neither approach can be improved.
Where trading decisions are going
Three changes are visible. The first is that the line between the two approaches is fading: discretionary traders increasingly run their risk rules through software, and systematic traders increasingly use judgement to choose which rules run when. The second is AI: assistants can now turn a trader's description into rules in seconds, which makes it easier to become systematic and also easier to produce many untested strategies. Testing and honest review matter more as rules become cheaper to write. The third is regulation: since 1 April 2026 algos offered to retail traders in India run through brokers under exchange rules, with each strategy registered, which adds a layer of structure to automated trading.
Start with 500 free tokens
Connect your broker, build a strategy and test it on paper before a single live order goes out.
Frequently asked questions
Is systematic trading better than discretionary trading?
Neither is better in all cases. Systematic trading is more consistent and testable; discretionary trading is more flexible. Many traders combine them, using judgement to choose strategies and rules to run them and to control risk.
Can discretionary traders use automation?
Yes. Automating stops, position sizes and daily limits helps any trader, even if entries are chosen by judgement.
Do systematic strategies stop working?
They can, when market conditions change. Regular review, testing on new data and watching how live results compare with the backtest help detect when a strategy no longer works.
What is the biggest mistake systematic traders make?
Overfitting: tuning rules to past data until they look perfect, then trusting the result.
Can a rule-based strategy still use my judgement?
Yes, at the level of choosing. You can decide which strategies to run in which conditions, when to pause one before a big event, and how much capital each one gets. What the rules take over is the trade itself: the entry, the stop, the exit and the size, so that those decisions are not made under pressure.
How do I know if my judgement adds value?
Write down every discretionary decision with its reason, then compare your results with what a simple rule would have done on the same days. If the rule does as well, your judgement is not adding much.
Related guides
- Algo vs manual trading: the practical differences in India
- Algo trading risks: what can go wrong with automation
- What is algo trading: how automated trading works in India
References
- Michael W. Covel, *The Complete TurtleTrader*, HarperCollins (2007), on Richard Donchian's rules and the 1983 Turtle experiment
- D. Kahneman and A. Tversky, "Prospect Theory: An Analysis of Decision under Risk", *Econometrica* 47(2), 263–291 (1979)
- H. Shefrin and M. Statman, "The Disposition to Sell Winners Too Early and Ride Losers Too Long", *Journal of Finance* 40(3), 777–790 (1985)
- T. Odean, "Are Investors Reluctant to Realize Their Losses?", *Journal of Finance* 53(5), 1775–1798 (1998)
- B. M. Barber and T. Odean, "Trading Is Hazardous to Your Wealth", *Journal of Finance* 55(2), 773–806 (2000)
- SEBI, circular on safer participation of retail investors in algorithmic trading, 4 February 2025, in force for every broker from 1 April 2026 (sebi.gov.in)
- SEBI, study of profit and loss of individual traders in equity F&O, 25 January 2023 (sebi.gov.in)
- SEBI, study analysing profit and loss of individual traders in equity F&O, FY22 to FY24, 23 September 2024 (sebi.gov.in)