Posted in My Daily Journal · 23 October 2026
The opening range is pinned as 31 one-minute bars judged from 09:46, futures fills are measured against replay, liquid stock options fail once more, a time-series foundation model finds nothing, a twelve-principle audit corrects four of them, ten independent reviewers vote on three proposals, and BankNifty at one lot goes back to paper.
Exactly how many bars are in the opening range?
Thirty-one: 09:15 through 09:45 inclusive, judged from the 09:46 bar onwards, and a re-run under that convention confirms the earlier results stand. It sounds trivial, but a one-bar disagreement between the simulator and the bot is exactly the kind of thing that quietly makes a paper record worthless.
How do live fills compare with replay?
On the pivot futures leg, live fills came in 10.6 points better than replay while paper fills came in 20.2 points worse, which says the paper feed, not the market, is the pessimist. Replay makes ₹465 per trade, about ₹13.2 lakh over twelve years, with a worst drawdown of ₹3.52 lakh.
Liquid stock options, one more time?
The door nets ₹226 per trade and the Williams buyer loses ₹283 per trade with a day-clustered t of −2.46, significantly negative, so it is out. A time-series foundation model was also pointed at the index: 50.9% directional accuracy, and a correlation of 0.15 with the day's range against 0.44 for simply using the opening-range width, so nothing there either.
November–December at ten lots looks fine on average. Is that reliable?
+₹7.19 lakh in total, 9 of 11 years positive, mean +₹65k, median +₹41k — and 2023 lost ₹2.04 lakh, so the median and the worst year are the numbers to plan around, not the mean. The same session audited twelve working principles and corrected four: premium-based exits sit on a plateau rather than a peak, the fourth quarter is the weakest rather than the strongest, drawdown must be measured against cash rather than against equity, and a 15:15 versus 15:25 exit makes no difference. Of the 100-plus cells tested, the keepers are skip-the-filled-gap, the 12:00 cutoff and the 31-bar range.
Is there a strike that is clearly best, or a market state with 80–90% odds?
The strike is a plateau: ITM-100, ITM-150 and ITM-200 give ₹349, ₹415 and ₹368 per trade, so ITM-150 stays, but nothing turns on the choice. No market state tested gives 80–90% odds; the best is 45–46%, barely above the base rate. Lot tiers by confidence fail, and pyramiding into a winner is worse than a single entry.
And the ten reviewers?
Ten independent reviewers — separate sessions given the same data and the same pre-registered pass bar — voted on three proposals. Skipping score-zero days: shadow-log 9, adopt 1. Confidence-based lot tiers: reject 10 of 10. A strike change: no change 10 of 10. BankNifty at one lot got a 4–1 no on top of that: since 2023 it loses ₹250 per trade, and paper fills ran ₹507 per trade worse than the backtest, so it goes back to paper.