Paper trading is only useful if you treat it as a test with a pass mark you set in advance. Run without one and you will interpret whatever happens as encouraging.
Before you start: set the pass mark
Write these down now, before the first trade. Deciding afterwards is how people talk themselves into going live.
- How many days you will run — two weeks minimum
- What result would make you go live
- What result would send you back to the drawing board
- What result would mean the strategy is wrong, not just unlucky
- That you will not change parameters mid-run
The setup
1
Backtest first
Paper trading is slow. Do not spend two weeks on something a backtest would have ruled out in two minutes.
2
Build the strategy exactly as you intend to run it
Same times, same strikes, same lot size, same stop. Not a simplified version.
3
Deploy it in paper mode
On Arthalab this is a mode on the same engine, so the behaviour matches live.
4
Start it
Deploying is not starting. This catches people repeatedly.
5
Confirm it evaluated on day one
Check the logs after your entry time rather than assuming.
Deploy and start are separate actions, and a paper strategy that was never started produces a very calm two weeks that proves nothing.
The daily check
Five minutes after market close. Not during the day — watching tick by tick is how you end up interfering.
| Check | What you are looking for |
|---|---|
| Did it enter? | If not, why — the logs will say |
| At what price, and at what time? | Against what you expected |
| Did it exit as configured? | Scheduled exit, stop, or target |
| Day's P&L | Recorded, not judged |
| Anything surprising? | This is the valuable column |
The logs are the point of the daily check. P&L tells you the outcome; the logs tell you whether the strategy did what you designed.
What the number is for
Record the daily P&L without reacting to it. Two weeks is far too short a sample for the number to mean anything on its own — you are testing behaviour, not profitability.
What the first week usually finds
Almost always configuration rather than strategy, which is exactly what you want this stage for. The usual list, roughly in order of frequency:
| Finding | Usually because |
|---|---|
| It never entered | The strategy was deployed but never started |
| It entered at the wrong strike | The strike rule was not what you thought it was |
| It never exited | The exit time was set outside market hours |
| The stop never fired | It was configured on the wrong leg |
| It entered twice | A duplicate deployment left running |
Every one of these costs nothing to find in paper mode and real money to find live. The logs identify all five in under a minute, which is why the daily check is reading the logs rather than reading the P&L.
What paper trading can prove
- The strategy evaluates when it should. Timing, conditions, no silent skips.
- Strike selection behaves on real live prices rather than historical ones.
- Exits fire correctly, including the stop and the hard square-off.
- Your configuration is right. The most common real finding.
What it cannot prove
- Real fills. Simulated fills do not pay the spread the way your order will.
- Broker behaviour. Rejections, margin checks and order type quirks are absent.
- Profitability. Ten trading days is not a sample.
- How you handle a real loss. This is the biggest gap and the reason the next step is live-small.
After two weeks
1
Compare against the pass mark you wrote down
The one from before you started.
2
Compare behaviour against the backtest
Not the P&L — whether it entered and exited where you expected.
3
List every surprise
Each one is either a bug in your configuration or a gap in your understanding.
4
Fix configuration issues, then re-run
A fresh two weeks. The previous run tested a different setup.
5
If clean, go live with one lot
The gap between paper and live is the next thing to measure.
The short version
- Write the pass mark down before you start
- Backtest first — do not spend two weeks on something a backtest would reject
- Build it exactly as you intend to run it, then deploy and start it
- Check the logs daily after close; record P&L without reacting
- Changing a parameter mid-run restarts the two weeks
Frequently asked questions
Two weeks minimum, a month is better because it covers a full expiry cycle. Set the duration before you start rather than deciding as you go.
Yes. Paper trading is slow, so do not spend two weeks on a strategy a backtest would have ruled out in two minutes.
You can, but it restarts the clock. After a change you are testing a different strategy on data you have already seen, so the earlier days no longer count.
That the strategy evaluates when it should, selects strikes sensibly on live prices, exits correctly, and that your configuration is right. The last one is the most common real finding.
Real fills, broker behaviour, profitability over any meaningful sample, and how you react to a losing day with real money. That last gap is why the next step is live with one lot.
No. Check the logs five minutes after close. Watching tick by tick is how you end up interfering with a test you set up to run itself.
Start with a free 3-day trial
Build a strategy, backtest it and run it on paper — no broker, no IP and no money needed to try it.
