proof #003NIFTY · BANK NIFTY · FINNIFTY (+ BTC, ETH)11 october 2026
Dinesh Kirola (Stock Burner) said:
Stock Burner, YouTube, 6 Oct 2024
His 9-20 EMA setup. Every minute of NIFTY, BANK NIFTY and FINNIFTY since 2015. We wrote his rules down before running the test, then counted how often his 1:3 target was hit — with real charges.
If the rules do what he says, how many trades should win?
Accuracy = his 1:3 target reached before the stop, on the 5-minute spot chart — the way he measures it. Statutory option charges and slippage included.
These are the actual files that produced the test. Open a step to see the implementation.
def load_market(sym: str, source: str = "kaggle", start: str | None = None, end: str | None = None) -> Market: kind = "crypto" if sym in CRYPTO else "nse" if kind == "crypto": f = RAW / f"{sym}_m1.npz" source = "binance" else: f = RAW / (f"{sym}_m1.npz" if source == "kaggle" else f"{sym}_hf_m1.npz") z = np.load(f) t = z["t"] keep = np.ones(len(t), bool) if start: keep &= t >= int(pd.Timestamp(start, tz=IST).timestamp()) if end: keep &= t < int((pd.Timestamp(end, tz=IST) + pd.Timedelta(days=1)).timestamp()) off = 19800 if kind == "nse" else 0 loc = t[keep] + offdef strat_nine20(m: Market, v: dict, rules: dict, zero: bool = False, optimistic: bool = False) -> tuple[list[dict], dict]: """The 9-20 EMA setup on `tf_min` candles (sb_rules.json -> nine20 for every rule's source). Bias: long only while EMA9 > EMA20, short only while EMA9 < EMA20. A 9/20 crossover arms the setup. range_filter='breakout': the cross must come with a range breakout in the same direction = a candle close beyond the extreme of the previous `range_n` candles, within `breakout_window` candles before the cross or after it (before the entry). [Ma1HnAXxww4: range -> breakout -> crossover] nochop: skip a cross that comes after >= 2 other crosses in the previous `chop_n` candles (EMAs intertwined). need_departure: after the cross, price must first leave the 9 EMA (a candle entirely beyond it) and then come back for the pullback. [uK7epk8-r9o 0:08:42 'निकला निकला निकला। और फर्स्ट टाइम पुलबैक'] max_entries_per_cross: 1 = first pullback only; null = every pullback while the trend lasts. Entry candle (long; short mirrored): low <= EMA9 (it comes to the 9 EMA) and close > max(EMA9, EMA20), and with entry='touch9_bull' a green candle. Entry at its close. Stop: 'entry_candle' low/high or 'swing5' = extreme of the last 5 candles. Target rr x risk (null = none). exit_ema20: exit at the close of a candle that closes beyond the 20 EMA against the trade.""" tf = v.get("tf_min", 5) cd = candles(m, tf)def cost_points(rules: dict, sym: str, date: str, spot: float, zero: bool = False) -> float: if not zero and sym in CRYPTO: # Delta Exchange perpetual futures: taker fee per side (+18% GST on the fee) + slippage per side, % of price K = rules["costs_crypto"] return spot * 2 * (K["taker_fee"] * (1 + K["gst"]) + K["slippage_pct_per_side"] * rules["costs"].get("slippage_mult", 1.0)) return _cost_points_nse(rules, sym, date, spot, zero)def _cost_points_nse(rules: dict, sym: str, date: str, spot: float, zero: bool = False) -> float: """Per-UNIT round-trip cost of an ATM option trade, in SPOT index points (premium points / delta): STT, NSE transaction charge (+GST), stamp duty, SEBI fee on an ATM premium, plus the stated slippage per side. Brokerage is a flat ₹20 per order whatever the quantity, so it is charged separately in rupees (brokerage_R).""" if zero: return 0.0 C = rules["costs"] gst = C["gst"]def bigbar_premise(m: Market, v: dict) -> dict: """'10 में से आठ बार ... बिग बार के बाद से डायरेक्शन होता है' (KpOwBYb3H7c 0:18:52): after a big bar, how often does price continue in its direction? Compared with ordinary candles of the same colour on the same side of the 9 EMA (the baseline).""" cd = candles(m, v.get("tf_min", 5)) C = cd["c"].to_numpy(); H = cd["h"].to_numpy(); L = cd["l"].to_numpy(); O = cd["o"].to_numpy() e9 = ema(C, 9) dirs, rng = bigbar_flags(cd, e9, v) day = cd["day"].to_numpy() n = len(cd) base = np.where((C > O) & (C > e9), 1, np.where((C < O) & (C < e9), -1, 0)) base[dirs != 0] = 0 def measure(sig): nc, ext, cnt = 0, 0, 0 for k in np.flatnonzero(sig):def cmd_record(a, rules: dict) -> None: """His PUBLIC RECORD: the Sensibull-verified daily P&L (Zerodha, 17 May 2023 - 4 Sep 2024) against his own claims for the same periods (record_claims.json), coverage of trading sessions, and market-data checks of the trades he showed. Writes sb_record_audit.json.""" C = json.loads(CLAIMS.read_text()) d = pd.read_csv(SENSIBULL_CSV) d["date"] = pd.to_datetime(d["date_ist"]) sess = _sessions() span = sess[(sess >= d["date"].min()) & (sess <= d["date"].max())] shared = set(d["date"]) unshared = [x for x in span if x not in shared] runs, cur = [], [] for x in span: if x in shared: if cur: runs.append(cur)His exact 9-20 rules: a range breakout, a 9/20 EMA cross, then a candle that touches the 9 EMA and closes beyond both EMAs; stop at that candle; target 1:3 on the spot chart. We also tested all his conditions, the 1-hour chart, a generous 1:1 target and riding the trend to a 20 EMA close.
Years at 60%+: none, in any index (0 of 36 index-years for his exact rules). Every 5-minute reading lost money after charges.
His verified record: 64 of 322 trading sessions between 2023-05-17 and 2024-09-04 were never shared — including 3 Aug 2023, the “+5 lakh, 150% ROI” vlog day Moneycontrol reported on. Sharing stopped on 4 Sep 2024.
August 2023 (Moneycontrol, 23 & 25 Aug): he was accused of editing a P&L to show about ₹5.6 lakh when it “should really have been” about ₹4.3 lakh; Angel One advised its clients not to engage with him; Zerodha removed his referral link. His apology video was later retitled “…Not a Scammer” and is now private.
His 2025–26 crypto results (“10 lakh into 2 crore in six months”) are shown as his own screen recordings. We found no third-party verified record for them.
“Marketing accounts” are an industry practice insiders describe (and an undercover investigation reported in Aug 2026). We make no claim about his account.
Full method and limitations ↗One version at a time. Select another reading to see what changes. Every losing version stays in the table.
His exact rules, all his conditions, his “most accurate” 1-hour chart, a generous 1:1 target, and riding to a 20 EMA close.
5,782 trades · 29.9% net-positive · -0.214R per trade · PF 0.76
Curve = cumulative R, not dollars or a compounded account. Prepared from the downloadable CSV, with bucket extrema retained. Win-rate reference uses a separate percentage scale.
| Receipt | |||||
|---|---|---|---|---|---|
| 522 | 33.5% | 0.095 | 1.121 | CSV ↓ | |
| 6,217 | 23.8% | -0.154 | 0.803 | CSV ↓ | |
| 2,648 | 27% | -0.225 | 0.743 | CSV ↓ | |
| 5,819 | 28.6% | -0.194 | 0.768 | CSV ↓ | |
| 6,075 | 48.9% | -0.214 | 0.642 | CSV ↓ | |
| 506 | 32% | 0.005 | 1.006 | CSV ↓ |
Yearly totals in R. Green above zero; red below. The last year is partial (data ends 2026-05-14). The raw JSON keeps every year’s win rate and trade count.
Exact counts from this CSV. Longest consecutive stop-loss run: 20. A day-end loss is not counted as a stop.
His most-viewed strategy video: rejection at the 9 EMA, 1:2 target or exit on a close across the 9 EMA.
8,031 trades · 36.4% net-positive · -0.323R per trade · PF 0.521
Curve = cumulative R, not dollars or a compounded account. Prepared from the downloadable CSV, with bucket extrema retained. Win-rate reference uses a separate percentage scale.
| Receipt | |||||
|---|---|---|---|---|---|
| 8,044 | 36% | -0.262 | 0.576 | CSV ↓ | |
| 8,024 | 34.7% | -0.377 | 0.467 | CSV ↓ | |
| 8,031 | 36.4% | -0.323 | 0.521 | CSV ↓ |
Yearly totals in R. Green above zero; red below. The last year is partial (data ends 2026-05-14). The raw JSON keeps every year’s win rate and trade count.
Exact counts from this CSV. Longest consecutive stop-loss run: 9. A day-end loss is not counted as a stop.
A spot-chart stand-in for his Big Bar entries (his option-chart condition can’t be coded).
201 trades · 33.3% net-positive · -0.319R per trade · PF 0.445
Curve = cumulative R, not dollars or a compounded account. Prepared from the downloadable CSV, with bucket extrema retained. Win-rate reference uses a separate percentage scale.
| Receipt | |||||
|---|---|---|---|---|---|
| 6 | 33.3% | -0.528 | 0.342 | CSV ↓ | |
| 6 | 33.3% | -0.528 | 0.342 | CSV ↓ | |
| 6 | 50% | -0.458 | 0.36 | CSV ↓ | |
| 662 | 34.3% | -0.244 | 0.462 | CSV ↓ | |
| 12 | 33.3% | -0.217 | 0.642 | CSV ↓ | |
| 12 | 33.3% | -0.274 | 0.548 | CSV ↓ |
Yearly totals in R. Green above zero; red below. The last year is partial (data ends 2025-12-29). The raw JSON keeps every year’s win rate and trade count.
Exact counts from this CSV. Longest consecutive stop-loss run: 3. A day-end loss is not counted as a stop.
The same idea on BTC and ETH (Binance prices standing in for Delta Exchange), 1-hour and 15-minute.
1,601 trades · 26.2% net-positive · -0.197R per trade · PF 0.785
Curve = cumulative R, not dollars or a compounded account. Prepared from the downloadable CSV, with bucket extrema retained. Win-rate reference uses a separate percentage scale.
| Receipt | |||||
|---|---|---|---|---|---|
| 6,801 | 24.9% | -0.506 | 0.551 | CSV ↓ | |
| 1,251 | 27.3% | -0.135 | 0.848 | CSV ↓ | |
| 2,279 | 35.3% | -0.118 | 0.844 | CSV ↓ | |
| 2,034 | 50% | -0.233 | 0.623 | CSV ↓ | |
| 1,601 | 26.2% | -0.197 | 0.785 | CSV ↓ | |
| 6,804 | 26.2% | -0.345 | 0.664 | CSV ↓ |
Yearly totals in R. Green above zero; red below. The last year is partial (data ends 2026-09-23). The raw JSON keeps every year’s win rate and trade count.
Exact counts from this CSV. Longest consecutive stop-loss run: 22. A day-end loss is not counted as a stop.
5-minute chart, his exact 9-20 rules, 1:3 target, Jan 2015–May 2026 (5,782, 5,819, 5,821 trades).
“Accuracy… 60 to 70%, highest on the 1-hour” (19 Nov 2025). Same rules, 1-hour candles, 1:3 target.
“383 trades… 110 trades… 29% win rate” — read out by him on 4 Sep 2026 (his own company’s tool). He added it was still profitable: +58.7% with a −60.6% drawdown.
NIFTY 5-minute, our coding of his big-bar definition (BANK NIFTY and FINNIFTY have too few big bars under it). A coin toss either way.
Sensibull-verified Zerodha P&L, 2023-05-17 – 2024-09-04, against the NSE trading calendar. 3 Aug 2023 is not in the record.
Short source excerpts, not our paraphrase. Hinglish quotations are translated; timestamps let you hear the original. Re-uploads are labelled.
“In the last six months I turned my 10 lakh trading account into 2 crore.”
Stock Burner, YouTube, 17 Jan 2026
Watch on YouTube ↗ 0:04YouTube loads only after you press play. A source may disable embedding or become unavailable; the original timestamp link remains available.
What is promoted, what is promised, and what the primary sources actually say. These are published-source observations, not findings about anyone’s private accounts.
His site and video descriptions have carried referral links for Zerodha, Upstox, Angel One, Fyers, Alice Blue, Dhan, Delta Exchange and others. Moneycontrol reported in 2023 that he told his Telegram channel he earns a share of the brokerage fee from them.
Free only if you open a Delta Exchange account through his link and submit your Delta UID; otherwise ₹4,236–15,253 + GST. Delta’s own terms say its products are not offered through a SEBI-registered intermediary.
₹4,000–20,000. His terms: “No refunds… under any circumstances” and “We do not guarantee any income”. His site: “Please note that we are not registered with SEBI.”
Primary-source pages checked October 2026. Their contents can change; the receipts record what this audit used.
Dinesh Kirola (Stock Burner) or their team are welcome to respond. We will publish the response here. Found a mistake in a quote, rule or trade? Send the source and the row; we will check it and log the correction publicly.
Send evidence or a response ↗Corrections so far: none (11 october 2026).
The rules. The code. Every trade. Browse small previews here, or take the full research pack with you.
Public receipts · published 2026-10-11
realshyt proof data pack: https://realshyt.com/proof/stock-burner-dinesh-kirola-9-20-ema-strategy-backtestWhat this is: our backtest of the strategies Dinesh Kirola ("Stock Burner") teaches — his 9-20 EMA setup, his 9 EMA scalpingand his Big Bar entries — on every minute of NIFTY, BANK NIFTY and FINNIFTY from 2015 to 2026 (plus BTC and ETH), with statutorycharges and slippage; and his public Sensibull-verified P&L record checked against the trading calendar. 40 readings with every trade.Our test, our method, our opinion. Not financial advice.Files trades/*.csv every trade of every published reading (R = result in units of risk, after costs; cost_pts in index points) results_*.json every number for every reading (incl. the larger 9 EMA readings not published as CSV), with yearly results premises.json his premises tested directly (big bars; 5-minute vs 1-hour; trend days) crosscheck.json the same rules on a second data source (2021–2026) — same conclusion options_NIFTY.json his exact rules priced on real NIFTY weekly option premiums (a sample of expiries; cross-check only) headline.json the headline readings quoted in the video his_public_record/ his Sensibull-verified daily P&L (as published by him), his own claims, and our coverage + market checks rules.json the exact rules and costs, written down before the first run (dated snapshots in the repo's prereg/) METHOD.md his rules with verbatim quotes + timestamps, every ambiguity, what was not codeable, all results code/sb_backtest.py the Python used for everythingFound a mistake? Tell us on Instagram @realshyt__ and we will correct it publicly.Copy copies the visible page of source. Raw and Download contain the entire file.
realshyt proof data pack: https://realshyt.com/proof/stock-burner-dinesh-kirola-9-20-ema-strategy-backtest
What this is: our backtest of the strategies Dinesh Kirola ("Stock Burner") teaches — his 9-20 EMA setup, his 9 EMA scalping and his Big Bar entries — on every minute of NIFTY, BANK NIFTY and FINNIFTY from 2015 to 2026 (plus BTC and ETH), with statutory charges and slippage; and his public Sensibull-verified P&L record checked against the trading calendar. 40 readings with every trade. Our test, our method, our opinion. Not financial advice.
Files trades/*.csv every trade of every published reading (R = result in units of risk, after costs; cost_pts in index points) results_*.json every number for every reading (incl. the larger 9 EMA readings not published as CSV), with yearly results premises.json his premises tested directly (big bars; 5-minute vs 1-hour; trend days) crosscheck.json the same rules on a second data source (2021–2026) — same conclusion options_NIFTY.json his exact rules priced on real NIFTY weekly option premiums (a sample of expiries; cross-check only) headline.json the headline readings quoted in the video his_public_record/ his Sensibull-verified daily P&L (as published by him), his own claims, and our coverage + market checks rules.json the exact rules and costs, written down before the first run (dated snapshots in the repo's prereg/) METHOD.md his rules with verbatim quotes + timestamps, every ambiguity, what was not codeable, all results code/sb_backtest.py the Python used for everything
Found a mistake? Tell us on Instagram @realshyt__ and we will correct it publicly.
Open a 5-minute NIFTY (or BANK NIFTY / FINNIFTY) spot chart in IST and add the 9 and 20 EMAs. Find the recorded entry time, compare the stop and the 1:3 target, then follow the candles to the recorded exit. cost_pts is the per-trade cost in index points; rules.json lists the schedule.
The server samples uniformly from every trade in this version. The complete CSV stays out of your browser.
Times: UTC. Prices are CFD prices; costs can explain differences from the chart.
Open every trade in this version ↓A name is a start. A public claim with its source is better. Comment BURN on the audit reel and tell us what you want checked next.
Our test and our opinion of it. Short public excerpts used for commentary and review. Educational content, not financial advice. Past results do not predict future results.
All investigations ←