Squeeze ke baad breakout? Data kehta hai nahi

7 min read · data: 437-stock NSE universe · descriptive study, not advice

“Stock bahut din se ek range mein hai, volatility sukh gayi hai — ab phatega.” Yeh trading ki sabse purani kahaniyon mein se ek hai. Bollinger squeeze, NR7, coil, “calm before the storm” — naam alag, idea ek: jab price shant ho jaata hai, toh uske baad bada move aata hai.

Is kahani ke do hisse hain, aur humare data mein dono ka jawab ek jaisa nahi aaya:

1. Kya compression ke baad move kisi khaas direction mein aata hai? — Nahi.
2. Kya shant stock ke baad hi bada move aata hai? — Ulta: bade moves zyada tar pehle se garam (high-volatility) stocks se aate hain.

Yeh article dono ko alag karke dikhata hai. Yeh measurement hai, prediction nahi.

Pehle, “compression” ka matlab kya liya

Humne compression ko ek number se define kiya: 20-day realized volatility, us stock ki apni pichhle 1 saal ki volatility ke saath compare karke. Agar aaj ki volatility apne saal bhar ke range ke sabse neeche wale 25% mein hai, toh stock “compressed” hai. (Realized volatility = roz ke % moves kitne bade hain, uska average-type measure.)

Phir trigger: compressed state ke baad pehla din jab close 20-din ke high ke upar (ya low ke neeche) nikle, aur volume 20-din average ka kam se kam 1.5 guna ho. Yeh classic “squeeze breakout” hai. Direction humne predict nahi ki — breakout ne khud batayi.

Universe: 218 F&O stocks, 2016–2026. Entry agle din ke open pe (kyunki EOD signal dekh ke koi usi close pe entry nahi le sakta). Exit ek trailing rule se — stock apne highest close se 3×ATR neeche aaye toh bahar, ya 40 trading din baad.

Raw numbers achhe dikhte hain

Pehli nazar mein upside breakout wala row kaafi attractive lagta hai: +2.78% per trade, chhote losses, bade wins. Yahin pe zyada tar backtests ruk jaate hain. Hum yahin se shuru karte hain.

Asli test: random din pe entry lete toh kya hota?

Ek saaf sawaal: agar compression aur breakout ko hata do, aur bas random din pe wahi exit rule laga do, toh kya hota? Humne F&O universe ke har eligible din ko ek random entry maana (5 lakh se zyada simulated outcomes) aur same exit lagaya.

50.7th percentile matlab bilkul beech mein — random entry jitna hi. Woh +2.78% kahan se aaya? Do cheezon se:

• Market ka structural drift. 2016–2026 mein Indian F&O stocks lagbhag 28 trading din mein average +2.36% chadhe (raw, bina kisi signal ke).
• Exit rule. Trailing exit winners ko chalne deta hai aur losers ko jaldi kaat deta hai. Isse +0.4% aur skew aata hai.

Compression aur breakout ne kuch nahi joda. Downside breakouts toh seedha nuksaan the — kyunki ek upar jaate market mein short side hamesha drift ke khilaaf chalti hai.

Saal dar saal bhi dekha: 12 mein se 6 saal positive, 6 negative. Zyada tar profit sirf 2020 (+7.7%) aur 2023 (+4.5%) se aaya — dono bade V-recovery saal.

Ek doosra version bhi fail hua

Hamare paas ek aur rule tha: compression + futures OI direction (agar 5 din mein OI 5% se zyada badha toh long, ghata toh short). NIFTY 50 stocks, 2016–2026, 2,964 trades.

• Net per trade: −0.26%.
• Random entry bootstrap ke against: 4th percentile — yaani random se bhi kharab timing.
• 11 mein se koi bhi saal +0.4% per trade cross nahi kiya; 6 saal negative.

Toh “calm before the storm” ka kya hua?

Humne NIFTY 50 ke bade move episodes (2016–2026) ki forensics mein dekha ki episode shuru hone se pehle stock ki volatility kaisi thi. Result folklore ka ulta tha: bade episodes pehle se elevated volatility se paida hote hain — apni 20-day realized vol percentile ka AUC 0.65, aur VIX percentile ka 0.64 (0.5 = coin flip, 1.0 = perfect). Yaani “toofan” zyada tar tab aata hai jab hawa pehle se chal rahi ho, sannate ke baad nahi.

Aur yeh bhi sirf size ke baare mein hai. Woh bada move upar hoga ya neeche — yeh volatility nahi batati. Humari poori research ka ek saaf sabak yahi hai: market ka information magnitude mein hai, direction mein nahi. Volatility state risk aur position size sochne ke kaam aati hai; side chunne ke kaam nahi.

Kya dekhein (kya kharidein nahi)

• Compressed stock dekh ke “breakout aayega” socha ja sakta hai — lekin “upar ka breakout aayega” ka data support nahi karta.
• Agar koi backtest squeeze-breakout ka +2–3% dikhaye, ek sawaal poochiye: random entry with same exit kya deta? Zyada tar backtests yeh comparison nahi karte.
• /stocks/ pages pe volatility state dekhiye — woh batata hai stock shant hai ya nahi. Direction ka claim wahan nahi milega, jaan-boojh ke.

The limitation

Yeh study EOD data pe hai; intraday compression ya options IV-squeeze pe kuch nahi kehta. 2025–26 ka data research mein pehle use ho chuka tha, isliye usko decision ke liye nahi gina. Survivorship: current F&O list peeche lagayi gayi. Aur ek practical baat: yeh rule ek saath 50–117 positions maangta — normal capital ke liye feasible hi nahi. Base rates past describe karte hain; future predict nahi karte.

Methodology note. Squeeze-breakout test (Jul 2026): 218 F&O stocks (broad 486-stock robustness row reproduces it); setup = 20d realized-vol percentile < 25th vs own trailing 252 days, measured the bar before breakout; trigger = close beyond 20/55-day high/low with volume ≥ 1.5× 20d average; entry next open; exit 3×ATR(14) chandelier or 40-day time-stop; friction 0.30% cash / 0.08% futures round-trip; train 2016–20, validate 2021–24. Null 1 = random-entry longs with identical exit (505,359 outcomes, 1,000 draws). Compression + OI rule (Jul 2026): NIFTY 50, rv20 percentile < 0.30, 5-day total-futures-OI change ±5%, +12%/−6%/20-day exits, 1,000-draw bootstrap. Episode forensics: pre-episode feature AUCs for big-move episodes of either sign, NIFTY 50. Source: tietiy internal research (squeeze_breakout_setup_2026-07, compression_oi_rule_2026-07, nifty50_episode_forensics_2026-07).