Most strategies need you to predict whether prices go up or down. Statistical arbitrage doesn't. It bets on the 

 between two things staying true — and that's a very different, often calmer, way to trade. Here's how it works, in plain words.

Imagine two stocks that have moved together for years — same sector, same kind of business, rising and falling almost in step. Then one day, for no real reason, one of them shoots ahead while the other lags. A gap opens up between them.

If you believe that gap is temporary — that these two will fall back into step like they always have — you have the seed of a statistical arbitrage trade. You don't need to know if the market is going up or down. You only need the gap to close.

Statistical arbitrage — “stat arb” for short — is a 

market-neutral strategy that profits from temporary mispricings between related securities

, found using statistics rather than gut feel.

: any single trade might lose, but a sound edge applied across many trades comes out ahead on average. Second — and this matters — the word “arbitrage” oversells it. True arbitrage is risk-free. Stat arb is 

. It's a statistical bet that a historical relationship will hold. Usually it does. Sometimes it doesn't, and that's where the risk lives.

Stat arb spots two things that usually move together, waits for them to drift apart, and bets they'll come back together. It's a bet on a relationship, not on the market's direction.

Pairs trading: the whole idea in one move

, and once you get this, you get the concept.

Take two stocks that historically move together — say two private-sector banks. When the gap between them stretches unusually wide, you do two things at once:

Then you wait for the gap to close. When it does, you close both legs and pocket the difference. Because you're 

long one and short the other at the same time

, a general market move up or down largely cancels out — you're left with just the relative move between the two. That's what “market-neutral” means.

Buy the loser, short the winner, and bet the gap between them shrinks back to normal. Whether the whole market rises or falls barely matters.

 is just the gap between the two — often measured as a ratio (price of A ÷ price of B) or a difference. You watch this one number instead of two separate charts.

 means the spread tends to wander away from its average and then return to it. Stat arb is, at heart, a bet on mean reversion: “this gap is unusually wide, and wide gaps usually narrow.”

 (they move in the same direction day to day) without their gap being stable. What you really want is 

 — a deeper, lasting tie where the spread wanders but reliably comes home to an average. Correlation is a hint; cointegration is the thing you can actually trade.

How do you know the gap is “unusually” wide? The 

 tells you, by measuring how far the spread is from its average in standard deviations. A z of 0 means the spread is right at its average. A z of +2 means it's stretched two standard deviations wide — a typical entry trigger. You exit as it falls back toward 0.

Watch one number — the gap. The z-score tells you how stretched it is. Enter when it's very stretched, exit when it snaps back to normal.

. The gap is stretched. A looks rich relative to B.

 short A, long B — equal value on each side, say 

Short ~238 units of A at ₹840 · Long 500 units of B at ₹400.

 (820 ÷ 410 = 2.00). z ≈ 0, so you exit both legs.

 (before costs) — earned purely because the 

Here's the part that surprises people. The profit didn't come from the market going anywhere. To prove it, watch what happens if the whole market rallies and 

 stocks rise — but the gap still closes:

 A → ₹880, B → ₹440 (ratio back to 2.00).

Short A: (840 − 880) × 238 = −₹9,520  |  Long B: (440 − 400) × 500 = +₹20,000 → net 

Still a profit. The rally hurt the short and helped the long, but the convergence is what you captured.

Now a pure parallel move, no convergence:

 both rise 10% (A → ₹924, B → ₹440), ratio stays 2.10.

Short A: −₹19,992  |  Long B: +₹20,000 → net 

 convergence nets roughly zero. That's the proof: your profit comes only from the gap closing — the market's direction cancels out.

How a stat arb system works, step by step

Pairs trading is the doorway. The professionals scale the same idea to 

. A common version is cross-sectional mean reversion: rank a whole universe by how much each name has recently out- or under-performed, then 

short the recent leaders and buy the recent laggards

The logic is identical — bet on relationships reverting — just spread across many tiny, diversified positions so no single broken relationship can sink the book. This is the world large quant funds live in.

Stat arb is calm right up until a relationship you trusted stops being true. Respect that one risk above all others — size small and always keep a stop.

The biggest practical wrinkle in India is the short leg. You 

cannot freely short-sell stocks for delivery

 in the cash market — naked short selling is banned, and a cash short has to be squared off the same day (intraday only) unless you formally borrow the shares through the Securities Lending & Borrowing (SLB) system, which is fiddly and can get expensive.

So how do Indian traders run pairs cleanly? 

 The simplest, most common approach is to put 

 legs in the F&O segment — go long one stock's future and short the other's. A futures contract handles the short synthetically, with no borrowing and no same-day square-off pressure, so you can hold the position until the spread reverts.

In India, don't try to short the cash stock. Run both legs as stock futures — long one, short the other. That keeps the trade market-neutral and avoids all the short-delivery headaches.

That's the whole engine: find two things that belong together, wait for them to drift apart, and bet they'll reunite. Simple to say, demanding to do well — but once you see the market as a web of relationships rather than a row of up-or-down guesses, you start noticing these gaps everywhere.