Killing the Condor: What 11 Months of Backtest Said

Four Sensex weekly iron condors are in this journal. The book on them is red. After the fourth I stopped adding fresh discretion to the problem and did the thing I should have done before the first one: I backtested it properly, over a long window, on more than one platform, and let the number decide whether the strategy lives.

It does not. This post is the autopsy, and the replacement.

What I tested and where

I ran the weekly Sensex iron condor over 01 September 2025 to 31 July 2026 — eleven months, a full mix of trending, ranging and shocked markets — on two independent backtesting platforms, StockMock and AlgoTest. Two platforms on purpose: if one has a data quirk or a fill assumption I do not like, the other is a check. When both agree, I trust the direction of the result even if I do not trust the last decimal.

Both agreed. Over that window the iron condor returned negative to marginal — somewhere around break-even at best, negative on realistic cost assumptions. Not a disaster, not a drawdown that blew up the account. Something worse in a way: a strategy that grinds, pays out its premium in small wins, and then hands most of it back on the moves it cannot survive, netting to roughly nothing for eleven months of screen time and margin.

The live results in this journal were not bad luck against a good strategy. They were a representative sample of a strategy that does not work well enough to justify running.

Why the condor underperforms

I do not need to speculate, because the structure explains itself and I have written about it four times already.

The condor's whole risk-reward is unfavourable by construction: on my sizing it risked around ₹3,630 to make a realistic ₹1,185 at a 50% target. That demands a win rate in the mid-70s just to break even. A short strangle has a strictly better risk profile at the point that actually matters, and here is the counter-intuitive reason: the wings I was buying for protection were costing me more than they were saving me.

On the condor I paid roughly two-thirds of my collected premium straight back out to buy the hedge legs. Those wings cap the disaster, yes — but over eleven months, disasters large enough for the wings to pay were rare, while the premium drag from buying them was constant. I was paying an insurance premium every single week against an event that showed up a handful of times, and the arithmetic of that trade, over a long enough sample, is negative. The hedge did its job on the worst days and quietly bled me on all the others.

What beat it

Over the identical window, on both platforms, a plain short strangle — sell an out-of-the-money call and an out-of-the-money put, no protective wings, managed with stop-losses instead — returned meaningfully more.

The mechanism is the mirror of the above. Without the wings, I keep the full premium on every quiet week, which is most weeks. I give up the structural cap on the disaster, but I replace it with a stop-loss on each leg — an active exit that, if it fires reliably, truncates the loss before it reaches the catastrophic zone the wings were there to cap. In effect I am swapping a paid, always-on hedge (the long options) for a free, conditional one (the stop). Over eleven months, in this data, the free conditional hedge won clearly.

I want to be careful about what that does and does not say. The strangle is not "safer" — it has a fatter tail than the condor, because a gap straight through a stop is a real and uncapped event. What the backtest says is that, priced across a long sample, the premium saved by not buying wings more than compensated for the extra risk taken on the tail. That is an empirical claim about this window, not a law of nature, and it lives or dies on whether the stops actually execute.

The strategy I am switching to

Retiring the condor, deploying this in its place. The rules, written down so they are testable and repeatable:

1. Entry: Monday, 09:45 am. Sell the ATM − 400 PE and the ATM + 400 CE of the Sensex weekly expiry. One lot each. No wings.
2. Stop-loss: 25% on each leg, independently. If a leg's premium rises 25% above where I sold it, that leg is closed. The two legs are managed separately — one can stop out while the other runs.
3. Exit: 15:00 on the day before expiry — the same never-hold-into-expiry-day rule the condor posts arrived at the hard way, now baked in from the start rather than learned through a ₹3,400 loss.

Two things about this design that come directly from the last month of mistakes.

The 25% stop is the whole strategy. The backtest edge exists only if the stops execute. Every condor post in this journal is, at bottom, a story about a stop that lived in my head and did not fire when it mattered. Moving to a strategy whose entire edge depends on stops firing, while my documented weakness is not firing stops, is either the smartest or the stupidest thing I have done — and the deciding factor is whether these go in as resting orders at entry, not manual decisions at 2 pm. If I run this strategy the way I ran the condors, it will underperform the condor.

The expiry-day exit is non-negotiable and pre-committed. The backtest assumed it. If I hold past it "just this once," I am no longer running the strategy I tested, and the backtest result stops applying to me.

What I am watching

1. Live vs backtest divergence on the stops. The backtest fills the 25% stop mechanically and instantly. Reality has slippage and gaps. The single biggest risk to this switch is that my live stop-execution is worse than the simulator's, which would quietly erase the edge. I will log every stop fill against its trigger and compare.
2. The tail I have chosen to accept. No wings means a gap through a stop is uncapped. I traded a known, paid, capped risk (the condor) for an unknown, conditional, uncapped one (the strangle) because the sample said the trade-off pays. The first real gap will test whether I actually believe that or just believed the backtest.
3. Sample honesty. Eleven months is a decent window but it is one regime-mix. I am not going to treat a good backtest as a guarantee. Ten live trades from now I will re-run the numbers including my real fills and decide again.

The condor is retired because the data said so, not because I got tired of losing on it — though I did. Letting the backtest overrule my attachment to a strategy I had already committed to publicly is the actual discipline here. The strangle has to earn its place the same way, and I have written the rules down so that in eleven months I can hold it to the identical standard.

Disclaimer

This is a personal trading journal. It is a record of my own trades, my own money and my own mistakes — nothing more.

Nothing here is a trade recommendation, a tip, a call, or advice of any kind. I am not a registered adviser and I am not qualified to tell anyone what to do with their capital. The strikes, premiums, entries and exits above are what I did, not what you should do. Backtested results are not live results and do not account for all real-world costs, slippage, or execution risk.

If you read this blog and place a trade because of it, that trade is yours. I accept no responsibility for anyone else's losses. Do your own research, size for your own risk, and understand that options can lose you more, faster, than you expect.