Understand smart order routing — Hinglish notes

6.5.5 · Stock-Market › HFT & Advanced Concepts

Smart Order Routing KYA hai?

YEH INPUTS KYUN? Kyunki screen pe "best price" ka matlab "best net outcome" nahi hota. Ek venue achhi price dikhaa sakta hai lekin bhaari fee charge kare, ya itna door ho ki jab tumhara order pahunche tab tak price gaayab ho jaaye.

Router decide KAISE karta hai? (First-principles derivation)

Hume woh venue (ya mix of venues) chahiye jo Q shares khareedne ki total effective cost minimize kare.

Step 1 — Effective cost per venue define karo

Ek buy order ke liye venue $i$ pe:

YEH FORM KYUN? Cost = shares ke liye jo pay kiya $+$ venue ne jo charge kiya $-$ venue ne jo wapas diya $+$ market kitna move hua. Har term ek real cash flow hai, isliye hum bas sum karte hain. Koi hidden constants nahi hain — yeh literally tumhara bank statement hai.

Step 2 — Fill probability account karo

Ek order jo achhi price pe baitha hai woh bekar hai agar kabhi execute hi na ho. Benefit ko probability $ρ_{i}$ se weight karo ki fill actually hogi:

Step 3 — Optimization

Child sizes $q_{1}, q_{2}, \dots, q_{n}$ choose karo taaki:

$max_{q_{1}, \dots, q_{n}} \sum_{i = 1}^{n} E_{i} (q_{i}) subject to \sum_{i} q_{i} = Q, 0 \leq q_{i} \leq L_{i}$

$L_{i}$ = venue $i$ pe available liquidity (displayed size)
Constraint $\sum q_{i} = Q$ ka matlab hai: poora parent order fill karo.

PRACTICE MEIN KAISE SOLVE HOTA HAI: router venues ko marginal effective cost ke hisaab se sort karta hai (sabse sasta pehle), phir greedily har venue ki available size $L_{i}$ tak fill karta hai jab tak $Q$ shares place na ho jaayein — yeh classic fill-cheapest-first (waterfall) strategy hai. Yeh optimal hai jab cost per share ek venue ke andar constant ho.

Routing styles

Worked Examples

Common Mistakes

Active Recall

Recall National Best Bid/Offer use karke ruk kyun nahi sakte?

Kyunki NBBO sirf displayed price dikhata hai. Real execution cost venue fees/rebates, latency (price pahunchne se pehle gaayab ho sakti hai), aur fill probability pe bhi depend karti hai. SOR net effective cost optimize karta hai, headline quote nahi.

Recall Effective buy price scratch se derive karo.

Cost per share = jo ask pay kiya $+$ taker fee $-$ rebate mila $+$ market-impact penalty. Cash flows sum karo: $P^{eff} = P^{ask} + f - r + impact$ .

Recall Sequential (waterfall) routing risky kab hota hai?

Jab liquidity fade ho — ek fast trader tumhara pehla child order dekhta hai aur tumhare baad wale child orders ke aane se pehle doosri venues pe quotes pull kar leta hai.

Recall Feynman: SOR ek 12-saal ke bachche ko explain karo

Tum 100 marbles khareedna chahte ho lekin kisi ek dukaan mein 100 nahi hain. Ek smart helper har dukaan ko ek saath check karta hai, dekhta hai kiski marbles sasti hain aur entry-fee kam hai, aur turant har dukaan ko sahi utni marbles-orders bhejta hai — taaki tumhe saari 100 kam se kam paise mein milin aur koi ek dukaan price nahi badhaaye.

Flashcards

Smart Order Router kya hai?

Ek automated engine jo ek parent order ko child orders mein split karta hai aur unhe venues pe route karta hai taaki best price, liquidity, aur net cost mile.

SOR exist kyun karta hai?

Kyunki modern markets fragmented hain — liquidity kai exchanges/dark pools/ECNs mein bikhari hui hai, aur koi ek venue ke paas sab kuch nahi hota.

Effective buy price formula

P^{e f f} = P^{a s k} + f - r + impact

(ask + taker fee − rebate + market impact).

Fill probability ρ se weight kyun karte hain?

Ek achhi resting price bekar hai agar kabhi execute na ho; expected value = ρ × edge × size.

Aggressive vs passive routing

Aggressive spread cross karta hai (fee pay karo, certain fast fill); passive rebate kamane ke liye post karta hai (uncertain fill, sasta).

Spray vs waterfall routing

Spray ek saath saari venues pe bhejta hai (fast, over-fill risk); waterfall venue-by-venue jaata hai (fee-saving, liquidity-fade risk).

Liquidity fading kya hai?

Sequential routing mein tumhare baad wale child orders ke aane se pehle doosri venues pe quotes pull ho jaana.

Maker–taker model

Venues takers se fee (

f

) charge karte hain aur makers ko rebate (

r

) pay karte hain taaki liquidity providers attract hon.

Passive posting mein adverse selection

Tum exactly tab fill hote ho jab market tumhare khilaf move karne wali hoti hai.

Waterfall optimality condition

Fill-cheapest-first optimal hai jab per-share cost har venue ke andar constant ho.

Connections

Market Microstructure
Order Book Dynamics
Maker-Taker Fee Model
Dark Pools
Latency & Colocation in HFT
Market Impact & Slippage
Best Execution Regulation (Reg NMS / MiFID II)
Execution Algorithms (VWAP, TWAP, POV)