Register
Tutorial

Practice a Strategy with QuantPower Virtual Trading

This tutorial runs a strategy in QuantPower's Forward test: live prices, test money, and no order sent to any broker. It covers switching an algo on in Forward test, placing trades by hand in the Forward test account, watching the first sessions, keeping a record, comparing the results with the backtest, and deciding when, and how small, to go live. Setting up takes ten minutes; a useful test takes several weeks.

QuantPower Research · Updated 15 Nov 2026 · 17 min read

Forward test runs the same algo engine as live trading. Orders, positions and MTM behave the way they would in a broker's book, but every fill happens in a separate account that starts with ₹10,00,000 of test money and never mixes with real money. Your real portfolio, funds and risk rules read your real brokers only. A backtest asks what the rules would have done on past days; Forward test asks what they do on the days in front of you, with the prices, gaps and quiet hours you cannot choose.

Before you start

  • A strategy you have backtested. Forward test is the second of three tests: backtest, then Forward test, then live with the smallest size. The tutorial on how to read backtest results covers the first.
  • Prices. With your own broker connected, Forward test uses live prices from your account. Without one, it uses the last close, not real time, which is enough to check the rules work but not to judge fills.
  • Tokens. An algo in Forward test costs 100 tokens per 28 days. Manual trades in the Forward test account cost nothing.
  • A plan for how long to test. Several weeks, across quiet and busy days, and at least a few of the days your strategy is built for, such as expiry days.

Two ways to use Forward test

WayWhat it isCost
An algo in Forward testYour algo switched on in Forward test instead of live. It enters, watches and exits by its rules with test money100 tokens per algo per 28 days
The Forward test account by handOrders you place yourself, filled at the price you see, in the same test accountNo tokens

Most traders use both: the algo in Forward test to check the strategy, and the account by hand to practise exits, adjustments and the order sheet without risk. Both use the same test account and its test money, so keep manual practice small while an algo is under test, or its results become harder to read.

A short history of trading on test money

Traders have practised on paper as long as markets have existed: write the trade down, follow the price, count the result. Exchanges and brokers later offered simulated accounts, and trading competitions on simulated money became common in the 1990s and 2000s. The weakness was always the same: a trade written down after the fact is filled at the price you wish you had got.

Forward testing an algorithm fixes that weakness. The rules are fixed before the day starts, the entries and exits happen when the rules say, and the fills are taken from live prices, not chosen afterwards. Professional system traders have long run new strategies "on the side" for weeks before giving them capital, a practice described in the trading-system literature as out-of-sample or walk-forward testing.1 In India, SEBI's 2024 circular on the use of real-time price data by platforms made the line between education, simulated trading and live trading clearer for everyone who offers such tools.2

Step 1: Switch an algo on in Forward test

  1. Open Algos → Option bots and open the algo you want to test.
  2. Tap Switch on in Forward test.
  3. Check the quantity. Use the same number of lots you would trade live. Testing 10 lots when you plan to trade 1 teaches the wrong lessons about fills, and so does the reverse.
  4. Confirm. The algo's card now shows that it runs in Forward test.

The same switch-on checks run as for a live algo, except the ones about a broker: kill switch and loss-limit checks for the real account do not apply to the test account. If the wallet cannot cover the 100 tokens, the switch-on is refused and no tokens are taken.

Step 2: Watch the first session

On the first day, check that the strategy behaves as written, not whether it makes money:

  • Entries happen at the time and on the conditions you set.
  • Strikes are the ones your rule should pick at that moment: at the money, a number of strikes away, by premium or by delta.
  • Stop-losses and targets are at the right distance from each entry price.
  • Exits happen at the target, the stop or the exit time.
  • The activity log on the card has one plain line for every entry, exit, re-entry and refusal.

An example: a Nifty short straddle algo sells the 25,500 call at ₹182.40 and the 25,500 put at ₹167.90 at 9:20, with a stop of 30% on each leg. The stops should sit at about ₹237.10 and ₹218.25, which are 30% above each entry price, rounded to the 0.05 tick. If the card shows anything different, stop the algo and fix the rule now. A rule that is wrong in Forward test costs nothing; the same rule live does not. Most first Forward tests find at least one setting that is not what the trader meant.

Step 3: Read the Forward test positions

Forward test positions appear on the chart with a line marked "Forward test", and the algo's card shows today's legs, their fills and the M2M. The Forward test account's fills also appear in the Trades list beside your brokers' fills, labelled as Forward test.

They do not appear in your real portfolio, funds or P&L. If you also trade live, check which account an algo runs in before you read its P&L as money.

Reading the result day by day

Each test day, read three figures from the algo's card, in this order:

  1. The day's M2M per lot. Write it in your record. Over a few weeks these figures become the Forward test's own equity curve.
  2. The worst moment of the day. How far the M2M fell before the day ended. A day that ended at +₹1,200 after being −₹7,000 at noon tells you more about the strategy than the final figure does.
  3. The exits. Which legs closed on their stop, which on their target and which at the exit time. Compare the mix with the backtest's exit reasons.

Resist the urge to judge the strategy after one or two days. A strategy that loses on its first two Forward test days may be fine; one that wins on both may not be. The comparison with the backtest, after enough days, is what counts.

Step 4: Trade by hand in the Forward test account

Use the Forward test account to practise what the algo cannot practise for you:

  • Exits under pressure. When a test position moves against you, practise exiting from the position line on the chart, or from the Scalper.
  • The order sheet. Order types, limit prices, and stop orders with a trigger and a limit.
  • Adjustments. Rolling a leg, adding a hedge, closing one side of a straddle.

Manual trades in the Forward test account fill at the price you see when the order is placed. They cost no tokens. They are a rehearsal, so treat them like real trades: same size, same rules, same discipline.

Step 5: Keep a record

Forward testing is only useful if you treat it as real:

  • Note every trade, and every time you wanted to break a rule. A trade journal makes the notes easy to review.
  • Do not reset after a loss. "Reset" clears the Forward test account and stops the algos running in it. A test loss is the cheapest lesson you will get; erasing it erases the lesson.
  • Do not change the rules mid-test. If you change something, note the date and start a new test period. Comparing a changed strategy with the old results tells you nothing.

Step 6: Compare with the backtest

After a few weeks, run a backtest over the same days the Forward test covered, and compare:

What you seeWhat it usually means
Forward test close to the backtestThe rules behave live as they did on past data
Forward test clearly worse, on the same daysFills, spreads or rules that act differently on live prices; find which
Forward test better than the backtestUsually luck over a short period; do not raise the size because of it
Many more trades than the backtestA rule triggers on live ticks that did not show on the backtest's candles
Fewer tradesAn entry condition depends on data the backtest saw differently, or the algo was refused entries; read the log

The gap is the useful part. A strategy that does in Forward test what it did in the backtest has passed its second test. A strategy with a gap needs an explanation before it gets real money.

Step 7: Know what Forward test cannot show

  • Your broker's checks. Forward test does not send orders to a broker, so margin refusals, broker risk blocks and exchange price bands do not apply. The first live run is where you meet them.
  • Fills in thin strikes. Test fills use the price you see; a real order in a far strike may move that price.
  • Your nerves. Losing test money is not the same as losing your own. Expect to feel live losses more, and decide your rules for a bad day before the first live trade, not during it.
  • A missed broker sign-in. Live algos need a signed-in broker every day; Forward test algos do not. Build the habit of signing in before you go live.

Step 8: Go live, small

When the Forward test has covered enough days and matches the backtest closely enough:

  1. Switch the algo off in Forward test.
  2. Switch it on live, with your broker, at the smallest size: one lot.
  3. Set Smart Exit and the daily loss limit for the account before the first live entry.
  4. Watch the first live sessions as closely as the first Forward test sessions.
  5. Increase the size only when the live results match the Forward test.

The first live weeks are a third test, not the finish line. Compare them with the Forward test for the same kinds of day, as you compared the Forward test with the backtest. If live results fall clearly short, the cause is usually something Forward test could not see: real fills, broker refusals or a missed sign-in.

Tokens carry over sensibly: an algo paid for as Forward test that goes live in the same 28 days pays only the difference, 100 more tokens; an algo paid for as live can run in Forward test in the same period for nothing more.

A worked example: four weeks of Forward test

A trader backtests a Nifty iron condor entered at 10:00 on Mondays and Wednesdays, with a stop on each short leg and an exit at 3:10 PM. The backtest over the last year shows a steady result with a worst day of −₹9,400 per lot.

Week 1. She switches the algo on in Forward test with one lot. On Monday it enters at 10:00, but the strikes it picks are one strike further out than she expected. She reads the rule: it picks by premium, nearest to ₹40, and premiums were high that morning. The rule is doing what it says. She keeps it.

Week 2. On Wednesday a short call is stopped out at 1:45 PM and the algo re-enters, because she had left re-entries at one. The re-entry is stopped too. The day ends at −₹6,100. She notes that the backtest showed re-entries rarely helped, and decides to set re-entries to zero, starting a new test period the following Monday.

Weeks 3 and 4. Four entries, three small gains and one loss of −₹3,200. She runs a backtest over the same four days: the backtest shows a total within ₹900 of the Forward test.

Decision. The rules behave as tested. She switches the algo on live with one lot, sets Smart Exit at −₹12,000 for the day, and plans to review after four more weeks.

A worked example: practising exits by hand

A trader new to options wants to practise managing a short strangle before risking money. In the Forward test account she sells one lot of a Nifty call and one lot of a put, each about 300 points from the price, at the prices on screen.

Over three days she practises: placing a stop order on each leg with a trigger and a limit; dragging the stop line on the chart to a new price and confirming in the order sheet; exiting one leg from its position line when it doubles; and adding a hedge when the market moves sharply. None of it costs tokens, and none of it reaches a broker. By the time she trades the same strategy live, the screens are familiar and her attention can go to the market instead.

Setting up the test to match live

A Forward test is only as useful as it is close to how you will trade live. Before you start, check five things:

  1. The same size. The lots you will trade live, not more and not fewer.
  2. The same days and times. If the strategy will run live on expiry days only, run it in Forward test on expiry days only.
  3. The same strike rule. Do not switch a premium rule to a fixed strike for the test because it is easier to read.
  4. The same prices. Connect the broker you will trade with, so the test uses the prices your account will see.
  5. The same limits. Note the account limits you will use live, such as Smart Exit's stop-loss, and check after each test day whether they would have fired.

The fifth point matters because the account controls (Smart Exit, the daily loss limit and the kill switch) cover your real brokers, not the Forward test account. Check them by hand against the test's daily M2M: "on 14 November the test lost ₹8,700; my live Smart Exit line is ₹8,000; it would have closed the day at about −₹8,000."

Signals and webhooks in Forward test

Signal algos and webhooks can use the Forward test account too:

  • A signal in auto mode can be pointed at your Forward test account instead of a broker. When it fires, it places the order in the test account, with every check that applies there. Any change to the signal's condition or order disarms it, so you must arm it again on purpose.
  • A webhook starts and stops your algos in the place each one ran before. An algo that ran in Forward test restarts in Forward test; a webhook hit never moves an algo from test money to real money.

This makes Forward test useful for strategies driven by Arth.live price alerts or alerts from other charting software: the alert, the webhook and the algo all run as they will live, and only the fills are in test money.

What to write in your record

A short record for every test day turns a few weeks of Forward test into evidence:

NoteExample
Date and kind of day18 Nov, gap-up then range
Entries and exits as expected?Yes; call leg stopped at 13:45
The day's result per lot−₹2,900
Would live limits have fired?No; Smart Exit line −₹8,000
Anything you wanted to overrideWanted to exit the put early at 14:30; did not
A rule questionIs a re-entry useful after a stop on expiry day?

The fourth and fifth columns are the most valuable. The first tells you whether the account limits suit the strategy; the second tells you how you are likely to behave live.

Deciding the live size

A Forward test that matches its backtest says the rules work as written. It does not say how much capital the strategy needs. Decide the live size from three figures:

  • The worst day in the backtest and the Forward test, per lot.
  • The longest losing streak in the backtest.
  • The account loss you can take in a day without changing your plans.

If the worst day was −₹9,400 per lot and your daily limit is ₹10,000, one lot is the size, whatever the strategy's average day looks like. Start there, and add a lot only after the live results have matched the Forward test for several weeks.

A worked example: a strategy driven by alerts

A trader uses an indicator rule on Arth.live: when the 15-minute Supertrend on Nifty turns up, she wants to buy one lot of the at-the-money call through an algo with a 25% stop and an exit at 3:10 PM.

  1. She builds the algo with one rule leg, "ATM call, weekly expiry", and switches it on in Forward test.
  2. She creates a webhook in Algos → Webhooks and links the algo to it.
  3. On the Arth.live chart she sets the Supertrend alert and links it to her webhook signal.
  4. Over three weeks the alert fires nine times. Each time, the webhook starts the algo in Forward test, the algo picks the at-the-money call at that moment and enters, and the stop or exit time closes it.
  5. She compares the nine trades with a backtest of the same rule on the same days. Seven match closely; two entries came a few minutes later in the Forward test, because the alert fires on a closed 15-minute candle while the backtest assumed the candle's close price.

The lesson is about timing, not about the rule: live entries will come at the price after the candle closes. She widens her expectation of slippage by a few points and keeps the strategy.

Forward test on expiry days

Expiry days need their own test, because options behave differently in the last hours: time decay is fastest, at-the-money options move sharply, and the margin rules for real accounts are stricter. A strategy built for expiry should collect several expiries in Forward test before it goes live. Since September 2025, NSE weekly expiries fall on Tuesday and BSE expiries on Thursday, so a Nifty expiry strategy and a Sensex expiry strategy each get one test day a week.

Two cautions specific to expiry. First, the extra 2% margin on short index options on expiry day applies to your real broker account, not to the test account, so a Forward test can run a size your live margin will not allow on expiry morning. Check the margin before you go live. Second, far strikes can trade at a rupee or less on expiry afternoon; a test fill at that price may not be available for a real order of any size.

Forward test for a manual trader

Forward test is not only for algos. A trader who trades by hand from the chart or the Scalper can use the Forward test account to practise a new style for a few weeks: scalping options, trading futures around VWAP, or managing a hedged position through a day. The rules for a useful manual test are the same as for an algo: the same size you would trade live, a record of every trade, no resets after a loss, and a decision at the end about whether the results justify real money.

How long to test

There is no fixed number of days, but some guides help:

  • Cover the days your strategy is built for. An expiry-day strategy needs several expiries; a strategy that trades twice a week needs several weeks.
  • Cover different kinds of day. At least one sharp move, one quiet range and one gap.
  • Collect enough trades to compare. Twenty or more trades make a comparison with the backtest meaningful; five do not.
  • Stop testing when the answer is clear, either way. A strategy that clearly does not behave as tested needs fixing, not more weeks.

For most strategies that trade a few times a week, four to eight weeks is enough to see whether the live behaviour matches the backtest. A strategy that trades once a month needs far longer, which is a reason to backtest it over a longer period before the Forward test, so the Forward test has less to prove.

Common mistakes

  • Testing a different size from the one you will trade live.
  • Resetting the account after a bad week.
  • Changing rules mid-test without starting a new period.
  • Reading a few good weeks as proof, and raising the live size because of them.
  • Forgetting the broker sign-in when you go live. Forward test never needed it.

Common problems

ProblemWhat to check
The algo did not enterRead the card's line: an entry condition not met, missing prices, or the wallet short of tokens at switch-on
Prices look oldWith no broker connected, Forward test uses the last close. Connect your broker for live prices
Forward test P&L shows in my real portfolioIt does not. Check which account the position belongs to: Forward test positions are marked
The algo restarted live after I restarted itIt cannot: a restart stays where the algo ran. Check whether it was switched on live separately
Reset stopped my algosReset clears the Forward test account and stops the algos running in it. Live algos are not touched

Where testing before trading is going

Testing on live prices before risking money has become standard practice for systematic traders, and the tools now make it cheap: an algo can run in Forward test with the same engine, the same rules and the same screens as live. As SEBI's framework for retail algos asks brokers to answer for every algo order, a record of how a strategy behaved before it went live is becoming part of trading responsibly, not only a personal habit.3

Next steps

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

Does Forward test use live prices?

Yes, when your own broker is connected: it uses the live prices your account sees. With no broker connected, it uses the last close, which is enough to check the rules but not to judge fills.

Do Forward test orders reach my broker?

No. Every Forward test order is filled in a separate test account that starts with ₹10,00,000 of test money. Nothing reaches a broker, and nothing mixes with your real funds or positions.

Can I run the same algo in Forward test and live?

Not at the same time. Switch it off in one place and on in the other. An algo paid for as live can run in Forward test in the same 28 days at no extra cost; going from Forward test to live pays only the difference.

Does Forward test check my broker's margin?

No. Forward test fills in a separate test account and never asks a broker, so it does not run your broker's margin, price-band or risk checks. Check the margin for your size before the first live run, especially for short options on expiry day.

What does Reset do to my Forward test account?

It clears the account back to its starting test money and stops the algos running in Forward test. Live algos and your real account are not touched.

References

  1. Robert Pardo, *Design, Testing and Optimization of Trading Systems* (1992), on out-of-sample and walk-forward testing
  2. SEBI, circular on the sharing of real-time price data with third parties, 24 May 2024 (sebi.gov.in)
  3. 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)
Forward test results use test money and do not guarantee live results. Trading involves substantial risk. This page is for information and is not investment advice.