How to Choose a Research Analyst Strategy

Sorting by win rate and deploying the top result is the worst available method, and it is what most people do. A filtering order that starts with constraints rather than with performance eliminates the majority of the list in under a minute and leaves you comparing things that are genuinely comparable.

Filter in this order

Minimum margin. If you cannot fund it, nothing else about it matters.
Maximum drawdown, converted to rupees at your size. If you would not sit through it, stop here.
Trade count in the published backtest. Below roughly a hundred, every ratio is provisional.
Whether you understand the rules. If you cannot explain it, you will abandon it during the first bad stretch.
Return over max drawdown. Only now, and only across what survived the first four.

Win rate does not appear on that list. It is the most prominent number on most listings and the least useful on its own.

What each published figure tells you

Figure	What it says	What it hides
Win rate	How often trades were profitable	The size of the losses
Risk and reward	Average win against average loss	How the two combine over time
Minimum margin	What the account needs for one lot	That it rises with volatility
Index	Which market it trades	Nothing — this one is straightforward
Backtest period	What conditions it was tested in	What conditions it was not

The third row catches people operationally. A published margin figure was computed at some point under some conditions, and both change. Treat it as a starting point and keep headroom above it.

Win rate and loss size, together

Two strategies can show the same expectancy and feel completely different to run.

	Strategy A	Strategy B
Win rate	85%	40%
Average win	Small	Large
Average loss	Large	Small
What it feels like	Frequent small wins, rare painful losses	Frequent small losses, occasional large wins
How people abandon it	After one bad day undoes months	After a long losing run

Most option-selling strategies look like A. Knowing that in advance is what lets you recognise a bad day as normal rather than as evidence the strategy has broken.

Converting drawdown to rupees

The single most useful thing you can do before deploying anything, and the step almost nobody takes.

Take the published maximum drawdown percentage

It is usually shown as a percentage of deployed capital.

Multiply by what you intend to deploy

Not your total capital — the amount going into this strategy.

Look at the rupee number

Would you keep starting this bot every morning while down that amount?

If the answer is no, reduce size or move on

Not later. Now, while you are calm.

Checking the backtest is current

An analyst can revise a published strategy. When the rules change, the strategy is flagged as having a pending backtest until it is refreshed.

If you see that flag, the published figures describe an earlier version of the rules. Judging the current strategy on them is comparing two different things, and a long-standing flag is itself informative.

Before you deploy

You can fund the margin with headroom for a volatility rise
You have converted the drawdown to rupees and accepted it
The trade count in the backtest is large enough to mean something
You can explain in one sentence what the strategy does
The backtest is current rather than flagged as pending
You have written your stop condition, in rupees, before deploying

Then paper trade it for at least two expiry cycles. It costs nothing beyond a plan and tells you how the strategy behaves now rather than how it behaved in the test period.

Running more than one

Three strategies run simultaneously on the Monthly Plan, with paper bots counting towards the same limit.

Worth knowing before you fill those slots: strategies on the same index expressing similar views are correlated. Three short-volatility strategies are not three independent bets — a sharp move hurts all of them at once, and the combined worst case is larger than any one of them suggests.

The short version

Filter by margin first, drawdown second, trade count third
Win rate without average loss alongside it is close to meaningless
Convert the drawdown to rupees at your size before anything else
A pending-backtest flag means the figures describe older rules
Paper trade before funding, and write your stop condition first
Three similar strategies on one index are not three independent bets

Frequently asked questions

Filter by what you can fund, then by whether the drawdown is acceptable in rupees at your size, then by whether the backtest has enough trades to mean anything. Compare performance only across what survives.

A high win rate paired with large average losses is worse than a moderate one with small losses. The win rate says nothing on its own about whether the strategy makes money.

What an account needs for one lot. It was computed at a point in time and margin rises with volatility, so treat it as a starting point and keep headroom above it.

No. It is the worst that happened in that sample. A longer or different period would very likely contain a worse one, so read it as a lower bound rather than a ceiling.

The analyst has changed the rules since the last backtest, so the published figures describe an earlier version. Wait for the refreshed numbers before judging it on them.

Three simultaneously, with paper bots counting towards the same limit. Strategies on the same index expressing similar views are correlated, so they are less diversified than they appear.

Yes. It costs nothing beyond a plan and shows how the strategy behaves now rather than how it behaved during the test period.

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.

Create a free account

All articles More in RA Strategies