Register
Learn guide

Compare Algorithmic and Manual Trading

Manual trading means a person watches the market and places each order. Algorithmic trading means software places orders according to rules decided in advance. Most traders who move from one to the other keep the same strategies; what changes is who executes them, how consistently, and at what cost in time and attention. This guide compares the two on execution, discipline, testing, risk, costs and the regulatory position in India, and sets out who is better served by each.

QuantPower Research · Updated 12 Nov 2026 · 18 min read

The choice is rarely all or nothing. Many traders automate the parts of their trading that follow fixed rules, such as a morning option strategy or the exits on every position, and keep the decisions that need judgement for themselves. The useful question is not "algo or manual?" but "which parts of my trading are rules, and which are judgement?"

Algo vs manual trading at a glance

Manual tradingAlgo trading
Who decides each tradeYou, in the momentRules you set in advance
Who places the orderYou, in your broker's appSoftware, through your broker's official API
Speed from signal to orderSeconds to minutesUsually under a second
ConsistencyVaries with attention and moodThe same every time the rule fires
Testing before risking moneyHard to test honestlyBacktest on past data, then Forward test on live prices
Time neededThe whole session for intraday tradingSetup, then a daily check
Main risksEmotional and inconsistent decisions, missed exitsWrong rules, rejected orders, sessions that have ended
FlexibilityAdapts to anything you noticeOnly what the rules describe
Rules in IndiaOrdinary broker rulesSEBI's retail algo framework, in force since 1 April 2026

A short history of how orders moved from people to machines

For most of the twentieth century, an order was a spoken or written instruction. In India, until the early 1990s, shares changed hands in the open outcry ring of the Bombay Stock Exchange, where brokers shouted bids and offers across a trading floor. Then electronic trading arrived. The NSE began screen-based trading in 1994, and the BSE moved from its trading ring to its BOLT screen system in 1995.1 Once an order was a message to a computer, it could also be written by one.

In the United States and Europe, firms began to send orders by program in the 1980s, and by the 2000s a large share of exchange volume came from algorithms. In India, SEBI allowed direct market access for institutional clients in 2008 and published its first broad guidelines on algorithmic trading on 30 March 2012, asking brokers and exchanges to put risk checks on every algo order.2 For the next decade, retail traders reached automation unofficially, through broker APIs and third-party platforms.

SEBI's study of individual F&O traders, published in September 2024, made the divide visible. It found that 93% of individual traders lost money over the three years to March 2024, while proprietary traders and foreign investors made most of their F&O profits through algorithmic trading.3 On 4 February 2025 SEBI issued its circular on safer participation of retail investors in algorithmic trading, which brought retail algos inside the broker's responsibility, asked exchanges to register algo providers and tied API access to a registered static IP. The framework came into force for every broker from 1 April 2026.4

Execution

Timeline of an iron condor's dozen orders placed by hand over a minute against software placing them in a second, hedge legs first (illustration)

A multi-leg option trade placed by hand takes a minute or more, during which prices move. An iron condor has four legs, each with a stop, plus an overall exit: a dozen decisions and orders. Software places the legs in a second, in the right order, with the bought legs first so that the broker sees a hedged position for margin before the sold legs arrive.

The difference matters most for exits. A stop that depends on a person seeing the price, deciding and tapping is slower than one the software watches every tick. In a fast market, a few seconds is the difference between a planned loss and a large one.

A worked example: entering an iron condor

Take a Nifty iron condor with one lot (65 units): sell the 24,800 put and the 25,200 call, and buy the 24,600 put and the 25,400 call as wings.

  • By hand: four orders in a broker app, typed one by one. If each takes 15 seconds, the last leg goes in a minute after the first. If Nifty moves 30 points in that minute and the sold legs each move by about ₹8 against you, the entry costs about 2 × 8 × 65 = ₹1,040 more than planned, before charges.
  • By algo: the four legs go in about a second, bought legs first. The slippage is the bid-ask spread on each leg, often ₹0.50 to ₹2 for strikes near the money: perhaps 4 × 1 × 65 = ₹260.

Neither figure is fixed. The point is that manual entry adds a cost that depends on how fast you type and how fast the market moves, while automated entry turns it into a small, predictable spread cost.

Discipline

Rules written in advance remove the decisions traders most often get wrong: holding losers, cutting winners early, trading more after a loss, and skipping a signal because the last one failed. Software does not get tired at 2:30 pm or anxious after three losses.

The discipline cuts both ways. An algo follows a rule that has stopped working just as faithfully as one that works, until you change it. Manual traders sometimes notice a change in the market before any rule would. Our guide to systematic vs discretionary trading covers the behavioural evidence on both sides.

A worked example: one month of a simple rule

A trader's written rule is: sell a Nifty straddle at 9:20 am on expiry day, with a 30% stop on each leg and an exit at 3:10 pm. Over four expiries, the rule would have taken four trades: two small wins, one full stop on both legs and one small loss.

Trading by hand, the same trader skipped the first expiry because "the market felt strong", moved the stop to 50% on the third because the loss "would surely reverse", and took the fourth trade twice the usual size to make up for it. The rule was the same. The results differ because the person changed the rule four times in a month. That gap is what automation removes.

A worked example: the cost of a missed exit

A trader buys one lot of a Nifty call at ₹120 with a stop at ₹85 and a target at ₹190. The planned loss is (120 − 85) × 65 = ₹2,275, and the planned gain is (190 − 120) × 65 = ₹4,550.

Trading by hand, the trader is in a meeting when the option falls through ₹85. By the time he looks, it is at ₹60, and he exits there: a loss of (120 − 60) × 65 = ₹3,900, about 70% more than planned. If that happens on one trade in five, a strategy that made sense on paper (win ₹4,550, lose ₹2,275) quietly becomes one that loses ₹3,900 too often to work.

With the stop watched by software, the exit goes as a market order as soon as the price touches ₹85. In a fast market it may fill a little lower, say at ₹83, a loss of ₹2,405. The gap between ₹2,405 and ₹3,900 on one trade is the whole case for automating exits, even for traders who keep their entries manual.

Testing

A rule can be tested on past data; a judgement cannot. Before an algo trades live, you can see how its rules would have behaved over many past days, what the drawdowns were and how many trades it took. A manual trader has only his own record, which is short, mixed with changing methods and hard to separate from luck.

Testing has traps of its own, covered in our guide to backtesting mistakes. A tested strategy is not a guaranteed one, but an untested strategy is a guess.

Testing in two stages helps:

  1. Backtest the rules on past data, including brokerage, taxes and slippage.
  2. Forward test the same rules on live prices with test money for a few weeks, and compare the results with the backtest for the same days.

Only then put real money behind the rules, at the smallest size.

Risk control

Both approaches can use stops. Only software can enforce them without exception, apply a daily loss limit across every position at once and act while you are away from the screen. A manual trader who promises himself to stop at a ₹10,000 loss for the day keeps that promise on most days, and breaks it on the days it matters most.

Algo trading adds operational risks that manual trading does not have: a rule with the wrong side or quantity, an order sent twice, a rejected leg that leaves a spread unhedged, a stale price, a broker session that ended overnight. Each has a control, described in our guide to algo trading risks.

RiskManual traderAlgo trader
Missing an exitCommon when away from the screenRare: the stop is watched every tick
Moving a stop "just this once"CommonOnly if the trader edits the rule
Fat-finger quantityPossible on every orderPossible once, in the rule, then repeated
Order sent twiceRarePossible without checks against duplicates
Losing several strategies on one dayLimited by attentionPossible unless an account-level limit exists

Time and attention

Manual intraday trading occupies the whole session, from 9:15 am to 3:30 pm in the cash market and 3:40 pm in F&O since 3 August 2026.5 A rule-based strategy needs a broker sign-in in the morning, a look at positions during the day and a review after the close. Traders with jobs, or who run several strategies at once, can usually do so only with automation.

The saving is real but not total. Someone still has to read the activity log, check that each leg filled, and decide when a strategy has stopped working. Automation moves the work from watching prices to watching the system.

Costs

Automated strategies tend to trade more often than manual ones, because they take every signal. Each trade pays brokerage, securities transaction tax and exchange charges, and STT on futures and options rose from 1 April 2026.6 A strategy's edge after costs must be checked before it runs.

A worked example: a manual trader who takes 8 Nifty option round trips a month pays about 8 × ₹40 = ₹320 in brokerage at ₹20 an order. An algo that takes every signal of the same rule might trade 25 times a month: about ₹1,000 in brokerage, plus STT, exchange charges, stamp duty and GST on each. If the extra 17 trades are no better than average, the higher trade count only raises the cost.

Software is a further cost. On QuantPower, charts, the option chain, the screener and the strategy builder are free, a backtest costs 2 tokens with the first 10 each week free, and a live algo costs 200 tokens (₹200 before GST) per 28 days, with orders unlimited inside it. Against this, a manual trader's cost is time, and the errors that come with it.

The regulatory position in India

Both are legal. Under SEBI's framework, algos offered to retail traders go through a SEBI-registered broker, with the provider empanelled by the exchange and each approved strategy given an identifier. A trader's own self-built algo that stays within 10 orders a second needs no registration, but its API access must come from a static IP registered with the broker, and broker API sessions end every day. Manual orders through a broker's app are unaffected.

The framework also separates strategies by whether their rules are visible. A strategy whose logic the trader can read counts as a white box. One whose logic is hidden counts as a black box, and its provider must be registered as a research analyst. Our guide to SEBI rules for option trading explains the framework in full.

Algo or manual, by kind of trader

The answer changes with what you trade and how often.

Option sellers. Short straddles, strangles and iron condors live and die by their exits: a stop on each leg, an overall stop for the position and an exit time. These are rules by nature, they need several legs placed in the right order, and the worst days come fast. Most option sellers who trade more than once a week gain from automating at least the exits.

Option buyers. Buying strategies depend on a quick move, and time decay punishes a stalled one. The entry is often a judgement about momentum, but the exit is a rule: a stop, a target, a trailing stop and a time limit. A hybrid suits many buyers: enter by hand or by a signal, exit by rule.

Intraday futures traders. Breakout, opening range and moving-average rules are easy to write and test. The danger is overtrading: an algo takes every signal, including the ones a human would skip in a dull market. Test the rule with costs, and add a filter or a cap on trades a day if the test shows many small losses.

Swing traders in shares. Trades last days or weeks, so speed matters less. The gain from automation is mostly in exits that wait for you: stop and target orders that stay at the exchange or with the broker until they trigger, such as a GTT order with two legs, one for the stop and one for the target.

Long-term investors. A monthly stock SIP or a yearly rebalance needs little automation beyond a reminder. The main risk is behavioural, selling in a fall, and a written plan helps more than software.

Part-time traders. Anyone who cannot watch the market through the session either automates or trades less. For them, automation is less about speed than about being present when the rule says to act.

What manual trading still does better

  • Unusual situations. A policy announcement, a trading halt or a freak trade may call for a judgement that no rule anticipated.
  • New ideas. A trader forms a view by watching the market; the rules come later.
  • Low-frequency decisions. A portfolio adjusted once a month gains little from automation.
  • Learning. Placing trades by hand teaches order types, margin and market behaviour in a way that watching an algo does not.
  • Reading the tape. Some traders' edge lies in reading the order book or the flow of news in the moment, which is hard to write as rules.

What algo trading does better

  • Rules followed every time, including the boring ones and the painful ones.
  • Speed and order of legs in multi-leg option trades.
  • Several strategies at once, across instruments and brokers.
  • Testing before risking money, on past data and live prices.
  • Exits while you are away, including account-level limits that no single strategy sees.

Who should automate

Automation suits traders who:

  • already follow written rules, or are willing to write them,
  • trade intraday or multi-leg options, where speed and the order of legs matter,
  • want to run more than one strategy,
  • cannot watch the market all day, or
  • know that their losses come from breaking their own rules.

It suits less well traders whose edge is reading news or order-book behaviour in the moment, and beginners who have not yet found a rule set worth automating.

A hybrid approach

Many traders keep both. Common splits are:

  • Automate the exits, keep the entries. Enter by hand when your view says so, and let an automated stop, target and exit time manage the trade.
  • Automate the routine, keep the judgement. Let the daily or expiry-day strategy run as an algo, and trade by hand only on the occasions that call for a view.
  • Automate the instruments, not the account. Let algos manage index options while you trade shares by hand, so the two never touch the same position.

The one rule for a hybrid is never to trade by hand in an instrument an algo is managing, because a manual trade changes the position the algo believes it holds.

Moving from manual to algo trading

  1. Write down the rules you already follow: instrument, entry, exit, stop, size and time window.
  2. Build them as an algo, from a template or your own legs, and backtest them on past data with costs included.
  3. Switch the algo on in Forward test for a few weeks, and compare its trades with the ones you would have taken by hand.
  4. Go live with one lot and a daily loss limit for the account.
  5. Keep trading by hand only in instruments the algo does not manage.

Our guide to algo trading covers the basics in more depth.

Which parts of your trading are rules?

Before choosing between algo and manual, sort your own decisions. Take your last 20 trades and, for each, write down why you entered, why you exited and how you chose the size. Then put each reason in one of three groups:

GroupExamplesWhat to do with it
Already a rule"Sold the straddle at 9:20 am on expiry day", "Stop at 30% of premium", "One lot per ₹2 lakh of capital"Automate it as it is
Could be a rule"Bought when the price broke the morning high", "Exited when the trend looked weak"Write it exactly: which high, which time window, what counts as weak. Then test it
Judgement"The budget speech sounded positive", "The order book looked heavy"Keep it manual, and keep it separate from the automated part

Most traders find that more of their trading belongs in the first two groups than they expected, and that the third group contains both their best and their worst trades. The exercise also shows where losses come from: if the worst trades all sit in the judgement group, the case for automating the rest is strong.

Common mistakes when moving from manual to algo trading

  • Automating a rule you never followed by hand. If you have not traded a rule manually or tested it, you do not know how it behaves. Test it first.
  • Going live at full size. The first weeks of a live algo find the operational problems: margins, rejections, sign-ins, slippage. Find them with one lot.
  • Keeping the manual habits. Editing the algo's stop in the middle of a losing trade brings back the very decision automation was meant to remove. Change rules after the close, not during the trade.
  • Running the algo and trading the same instrument by hand. The algo no longer knows what it holds.
  • Forgetting the account. Each algo has its stop, but nobody set a loss limit for the day across all of them. Set one before switching on the second algo.
  • Treating "automated" as "unattended". Sign in to your broker, read the log and check the positions every day.

Measuring whether automation helped

After three months of running a strategy as an algo, compare it honestly with the manual period before it. Four numbers are enough:

  1. Rule breaks. Count the trades in the manual period where you did not follow your own rule: skipped entries, moved stops, changed sizes. In the algo period the count should be close to zero.
  2. Average loss against planned loss. If your planned loss per trade was ₹2,000 and your average manual loss was ₹2,900, the gap is what late exits cost. Check the same figure for the algo.
  3. Slippage per trade. Compare fills with the prices your rule expected. Algo slippage should be small and steady; manual slippage tends to be larger on fast days.
  4. Cost per trade and trade count. If the algo takes many more trades than you did, check that the extra trades pay for their brokerage and taxes.

If the algo period shows fewer rule breaks and losses closer to plan, automation is doing its job even if the total for the quarter is lower, because three months of any strategy contain a lot of luck.

How QuantPower supports both

QuantPower is trading software, not an adviser: it gives you the screens to trade by hand and the engine to run your own rules, and it does not choose trades for you.7

  • Manual trading: the trading terminal shows positions, holdings and funds as your broker holds them; you can trade from the chart, from the option chain or from the Scalper, and every manual order opens the order sheet for your tap.
  • Algo trading: an algo is a set of legs with rules for entry, exit, target and stop-loss. A new algo is off; once you switch it on, live or in Forward test, it runs by its rules with no further tap. Templates are open samples you can change in every part.
  • Signals in between: a signal algo watches a condition on price, indicators or patterns and either prepares the order for one touch or, once you arm it, places it by itself. A signal that places its own order costs no tokens.
  • One set of limits for both: Smart Exit, the daily loss limit, a cap on orders a day and the kill switch cover manual orders and algos alike, across every connected broker.

Start with 500 free tokens

Connect your broker, build a strategy and test it on paper before a single live order goes out.

Where trading is going

The line between manual and algo trading is getting thinner. Regulation now treats a retail trader's API orders as algo orders with a registered address, so the route an order takes matters as much as who places it. AI assistants can turn a trader's description into rules in seconds, which makes the step from "what I do by hand" to "a rule I can test" much shorter. And account-level controls are moving into the platforms themselves, so a manual trader can now have the same daily loss limit that once needed a program. The judgement stays human; the execution keeps moving to software.

Frequently asked questions

Is algo trading better than manual trading?

For the same rules, automation executes faster and more consistently and can be tested first. Whether it is better for you depends on whether your edge can be written as rules. Judgement that cannot be written down cannot be automated.

Does automating a strategy make it earn more?

Not by itself. An automated bad strategy loses faster than a manual one. Automation adds consistency and speed to rules that already make sense after costs.

Can a beginner start with algo trading?

A beginner can start with open template strategies at the smallest size, in Forward test first. Building your own rules usually comes after some experience of how the market moves and how orders fill.

Can I trade manually and run an algo at the same time?

Yes, on the same account, as long as you do not trade by hand in the instruments an algo manages. A manual trade changes the position the algo expects.

Which is cheaper, algo or manual trading?

Manual trading has no software fee but costs time and the errors that come with it. Algo trading adds a software cost and often more trades, so check that the strategy's edge covers brokerage and taxes on every trade it will take.

What should I automate first?

The exits. Automated stops, targets and exit times on trades you still enter by hand remove the most expensive manual mistakes, and they teach you how automation behaves before you hand over the entries too.

References

  1. NSE, "About NSE: history" (screen-based trading from 1994), and BSE, history of BOLT (1995) (nseindia.com, bseindia.com)
  2. SEBI, broad guidelines on algorithmic trading, circular CIR/MRD/DP/09/2012, 30 March 2012 (sebi.gov.in)
  3. SEBI, "Analysis of profit and loss of individual traders dealing in equity F&O segment", 23 September 2024 (sebi.gov.in)
  4. 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)
  5. NSE, circular extending equity derivatives trading hours to 3:40 pm from 3 August 2026 (nseindia.com)
  6. Union Budget 2026-27: securities transaction tax on futures and options raised from 1 April 2026 (indiabudget.gov.in)
  7. QuantPower: manual order path, algo behaviour, signals and risk controls, from the app's code, checked 12 November 2026
Trading involves substantial risk. This page is for education and is not investment advice.