ICT/SMC strategy ko backtest kaise kare

AKQX QuantumTrade team ki taraf se · Sirf padhai ke liye, investment advice nahi. Examples ke numbers banawati hain aur tareeka samjhane ke liye hain, kisi asli result ke nahi.

Mukhya baatein

• Ye concepts discretionary hain aur alag logon ne alag tarah se define kiye hain, isliye definition tay kiye bina test nahi ho sakte.
• Test ka matlab: har judgement ki jagah ek number, aur use lambi history par mechanically lagana.
• Sabse bada dushman hindsight hai. Beet chuki har move me ek saaf pattern dikhta hai.
• Itni settings ke saath overfitting doosra bada khatra hai.
• Inme edge hai ya nahi, ye ek experiment ka sawal hai, aur jawab 'nahi' bhi ho sakta hai.

Ye ideas test karna mushkil kyun hai?

Inke aas-paas ki zyadatar sikhaai jaanboojh kar discretionary hai. "Price order block par laute, displacement ke saath, sahi session me, structure agree kare to..." ek vaakya hai jis par insaan act kar sakta hai aur computer nahi, kyunki har hisse me ek judgement chhupa hai: kaun sa block, kitna displacement, structure ki kiski definition.

Isliye log saalon in concepts par lage rehte hain aur kabhi nahi jaante ki unka version chalta hai ya nahi. Rules na ho to ginne ko kuch nahi, har haar ko misread aur har jeet ko confirmation bata diya jaata hai. Rules ko saaf karna concepts ko sahi sabit karne ke liye nahi hai. Ye data ko jawab dene ka mauka dene ke liye hai.

Shabd, aur har ek ke peeche chhupe faisle

• Order block: aam taur par bade move se pehle ki aakhri ulti candle. Faisle: body ya poori range, move kitna zordaar, return kab valid.
• Liquidity sweep: kisi saaf high ya low ke paar chhota sa move, phir andar wapas. Faisle: kaun se levels, kitna paar, kitni candles me wapas.
• Fair value gap: teen candles ka pattern jisme pehli aur teesri ki ranges overlap nahi karti. Faisle: gap ka minimum size, partial fill, kitni der valid.
• Break of structure: trend ki disha me pichle swing high ya low ke paar close. Faisle: swing ki length, wick ya close.
• SMT divergence: do correlated markets ka ek swing par ek-doosre se alag hona. Faisle: kaun si pair, swing ki definition.

Kisi concept ko rule kaise banayein

1. Definition ko numbers me likho. "Zordaar move" nahi, "3 candles me average candle range ka kam se kam 1.5 guna".
2. Use mechanical banao. Do log, ya ek hi insaan do din, wahi patterns mark karein, nahi to wo abhi bhi judgement hai.
3. Use causal banao. Rule sirf decision ke waqt tak ki jaankari use kare. Jis pattern ko pehchanne ke liye agli candle chahiye wo trade nahi ho sakta.
4. Entry, stop, target aur size jodo. Pattern strategy tabhi banta hai jab wo batata hai kya karna hai.

Do jaal

Hindsight: move ke baad zones banaoge to har rally me ek saaf order block milega. Pattern tabhi kaam ka hai jab koi rule use nateeje ke pata chalne se pehle pehchan sake.

Bahut saari settings: swing length, move strength, session windows, retest tolerance aur filters milke ek rule me dozen settings aa sakti hain. Har ek history ke noise me fit hone ka naya rasta hai. Rule chhota rakho, har setting thodi hilao, aur aise data par test karo jis par tune nahi kiya.

AKQX me kya hai

AKQX in ideas ko configurable blocks ki tarah deta hai taaki definition likhni hi pade: order blocks, liquidity levels aur sweeps, fair value gaps, structure breaks aur shifts, SMT divergence aur multi-timeframe confirmation jahan slower candle faster entry ko gate karti hai. Do AKQX-specific shabd bhi hain jinki apni working definition hai: PSP, yani do correlated pairs jinki candles ulti disha me close hoti hain, aur True Day Open, yani 00:00 UTC ka open jo ek daily anchor ki tarah kaam karta hai.

Result 350 me se score hota hai, aur negative expectancy, 1.0 se kam profit factor, 30 percent se zyada drawdown ya kamzor walk-forward, Monte Carlo ya overfitting score par apne aap reject ho jaata hai. Pass hona ek simulation ka description hai. Wo recommendation nahi, aur forecast bhi nahi.

Ek example, banawati numbers ke saath

Illustrative only (sirf samjhane ke liye): ek trader liquidity sweep ko aise define karta hai: price pichli 20 candles ke sabse bade high se kam se kam 0.2% upar jaye, phir do candles ke andar uske neeche close ho. Ek lambi history me ye rule 180 events dhoondta hai. Inme se 96 me agli paanch candles me price kam se kam sweep ke size jitna giri, aur 84 me nahi. Costs se pehle ye 53% hit rate hai, jo pehle se tay ki hui definition par hai. Ismein sahi nateeja ye hai ki 180 events par 53% coin flip ke itna kareeb hai ki akele costs use mita sakte hain. Agla kadam: 0.2% aur 20 ke aas-paas ke values dekho, aise period par check karo jis par rule chuna nahi gaya, aur fees aur slippage ghatao.

English me padhna chahein?

Yahi guide English me bhi hai: English version. Dono alag-alag likhe gaye hain, ek-doosre ka akshar-ba-akshar anuvaad nahi hain.

Questions about ict/smc backtest kaise kare

Aur padhein