OI Cohort Exits: Kaun Unwind Kar Raha Hai?

Zyadatar option-chain tools aapko batate hain ki open interest gira. Bahut kam ye batane ki koshish karte hain ki kaun ne use unwind kiya aur kyun. Nakshatra ka OI cohort-exit model uska signature feature hai: ek direction-aware inference jo ek OI drop ko todta hai aur estimate karta hai ki ye sellers the jo pain me wapas buy kar rahe the ya buyers jo liquidate kar rahe the, is breakdown ke saath ki wo positions kab likely khuli thi. Ye genuinely useful hai — aur ye ek estimate hai, jiske baare me ye page honest rehne ka khaas dhyan rakhta hai.

Ek bare OI drop ki problem

Jab kisi strike par open interest girta hai, kuch close hua. Par ek bare OI decline ambiguous hai. Kya wo writers (short side) the jo cover kar rahe the kyunki trade khilaaf chala gaya? Kya wo buyers (long side) the jo liquidate kar rahe the? Aur kya wo positions aaj khuli thi, ya dino purani leftovers hain jo bilkul alag cost bases carry kar rahi hain? Ek single number — “OI down 40,000” — inme se kisi ka jawaab nahi deta, phir bhi ye jawaab badal dete hain ki move ka matlab kya hai.

Exchange position holders publish nahi karta, to koi tool ise feed se padh nahi sakta. Jo ye kar sakta hai wo hai observable se dhyaan se reason karna: OI changes ka sequence, aur premium spot move akele se jo expected hota usse kaise alag move kiya. Wo reasoning hi cohort model formalise karta hai.

Cohorts: OI ko build hone ke time se group karna

Model pehle buildup ko IST calendar date ke hisaab se cohorts me group karta hai — roughly, “wo positions jo day X par khuli lagti hain.” Kyunki OI time ke saath accumulate hota hai aur Nakshatra har 5-minute snapshot store karta hai, ye fresh OI ko us din attribute kar sakta hai jis din wo dikha aur us cohort ko aage track kar sakta hai.

Ek position hai jise model date nahi kar sakta: wo open interest jo hamari data series shuru hone se pehle hi exist karta tha. Us OI ka koi observable entry day nahi, to use ek flagged carry-over cohort me rakha jaata hai. Ye ek deliberate honesty mechanism hai — un positions ke liye origin guess karne ke bajaye jise model ne kabhi open hote dekha nahi, model unhe carry-over label karta hai taaki aap jaano unka cost basis unknown hai.

Premium move ko delta-adjust karna

Model ka dil hai har OI drop ko classify karna ki position close hote waqt premium ne kaise behave kiya — par raw premium move misleading hai. Trending din par, ek option ka price zyadatar isliye move karta hai kyunki spot move hua, isliye nahi ki sentiment badla. Raw last-traded-price direction padhna isliye misfire karta hai.

To model delta-adjust karta hai. Strike ki apni implied volatility chain se use karke, ye ek Black-Scholes delta compute karta hai aur premium move ka spot-explained portion hata deta hai. Jo bachta hai wo us move ka hissa hai jo actual repricing reflect karta hai — wo signal jo hint deta hai ki kaun trade press kar raha tha. Phir har OI drop ko inme se ek me classify kiya jaata hai:

• short_covering_pressure — writers ke pain me wapas buy karne ke consistent.
• long_liquidation_or_mixed — buyers ke exit karne ke consistent, possibly muddied.
• mixed_unwind — evidence saaf ek taraf nahi dikhata.
• overnight_adjustment — session boundary ke paar gap-driven repositioning.

“Mixed” categories cop-out nahi hain; ye honestly use hoti hain jab bhi delta-adjusted evidence genuinely ambiguous hai, ek confident label force karne ke bajaye jise data support nahi karta.

Har cohort ke liye, model uske OI drops ki classification ko seller-pain ya buyer-pain se match karta hai aur ek 0 se 1 ke beech share report karta hai jo describe karta hai ki us cohort ka exit kaise breakdown hota hai. Output shares ke taur par present hota hai, raw unit counts nahi, to aap ise “is cohort ka unwind roughly 70% seller-pain dikhta hai” ki tarah padho, ek spuriously precise contract tally ki tarah nahi.

Aur ye ek confidence level ke saath aata hai, kyunki har step — cohort dating, delta adjustment, classification — observable data ke upar layered ek inference hai. High confidence matlab OI moves aur delta-adjusted repricing saaf align karte hain; low confidence matlab picture murky hai aur aapko label par halka hi lean karna chahiye. Honest framing essential hai: ye model ka sabse defensible estimate hai ki kaun exit kar raha hai, na ki position holders ka certified ledger.

Nakshatra me ye kaise dikhta hai

OI Cohort Exits panel Insights tab par rehta hai. Ek strike aur side chuno aur ye cohorts lay out karta hai — har entry-day group plus flagged carry-over cohort — har OI drop ki classification, seller-pain versus buyer-pain share, aur read ke peeche ka confidence ke saath. Option Chain page ek clickable notice dikhata hai jo yahan point karta hai, kyunki analysis baaki Insights signals ke saath belong karta hai. Ise wo samjho jo ye hai: kaun ek strike unwind kar raha hai uska ek transparent, direction-aware estimate — wo sawaal jise zyadatar tools attempt bhi nahi karte — apni uncertainty upfront bataye hue, chhupaye nahi.

Nakshatra tool me live dekho →