Chapter 4

The house always wins

Published 30 September 2026

A crore of people lost ₹1.81 lakh crore. That money did not evaporate. In a derivatives market every rupee lost is a rupee won by the counterparty, and the regulator's study names the counterparties. This is the chapter about who is sitting on the other side of your screen - and why it is not a person.

SEBI's study splits the market's profit and loss by category of participant, and the split is a confession. In FY24, proprietary traders earned about ₹33,000 crore in gross profits in the F&O segment, and foreign portfolio investors earned about ₹28,000 crore. Against that, individuals and others lost over ₹61,000 crore before transaction costs. Thirty-three thousand crore to the house's own desks. Twenty-eight thousand crore to foreign institutions. Sixty-one thousand crore out of the pockets of the crowd, in a single year, before the toll.

It is not a fair fight, and it was never meant to be

Who are these winners? 97% of the FPI profits and 96% of the proprietary profits in FY24 came from algo entities - firms that placed at least one algorithmic order in the year. Strip the polite definition and the picture is simple: the money being taken from retail traders is being taken by machines. Co-located servers in the exchange's own data centre, running models built by PhDs, reacting in microseconds, quoting both sides of every strike, harvesting the spread and the panic in equal measure. Your opponent is not another uncle with a view on Nifty. Your opponent is a rack of computers that never sleeps, never hesitates, and pays rent to the venue you are both trading on.

The marketing never shows you this counterparty. The app shows you a clean interface, a green button, and a chart. It does not show the server rack. It does not show that the price you just accepted was quoted by an algorithm that has traded this exact strike a million times, against a million people exactly as hopeful as you, and has priced your hope into the spread.

Ninety-nine point eight

The study records one more ratio that reframes everything: almost 99.8% of the participants in the F&O segment are individuals. The machines and the institutions are a sliver of the headcount. But the money flows the other way - the sliver takes ₹61,000 crore a year out of the crowd. This is not a market in the sense of a meeting of equals. It is a processing plant: a vast population of humans arrives with savings, and a small number of algorithms converts those savings into trading profits, in aggregate, every single year the regulator has counted.

That is why the industry's favourite defence - "trading is a skill, some people are better at it" - collapses on the study's own numbers. If this were a skill contest among peers, the winners would be other skilled humans and the ratio would drift with education and experience. Instead, year after year, roughly the same 91% lose, and roughly the same machines collect. The skill that wins is not available to the person the product is sold to. It runs on a server they will never see, owned by a firm they have never heard of.

The house edge, quantified

Every casino game has a house edge, and honest casinos are forced to publish it. The Indian F&O market's edge can now be read off a regulator's spreadsheet. In FY24 the crowd lost over ₹61,000 crore gross, and about ₹75,000 crore after costs. The professionals' gross take was about ₹61,000 crore - ₹33,000 crore to proprietary desks, ₹28,000 crore to foreign institutions. The residual - the difference between what the crowd lost net and what the professionals won gross - is the toll: roughly ₹14,000 crore in a year, split between brokers, exchanges, depositories and the exchequer. The game is zero-sum before costs and sharply negative-sum after them. The only guaranteed flows in the entire structure are the ones pointing away from the crowd.

The counterparty that never has a bad day

A human trader has a thesis, and a thesis can be wrong. The algorithms on the other side of this market mostly do not have theses at all - they have inventory and speed. They quote both sides of a strike, pocket the difference, hedge the residual in the underlying, and do it thousands of times a day. They do not need the market to rise or fall. They need it to move, and they need you to keep crossing their spread. The study's 96-97% figure is the proof that this works: the profits of the professional categories are not coming from brilliant directional calls. They are coming from a process, run at machine speed, against a crowd trading on a phone between meetings.

That process has no off day. It does not get scared after a gap-down, does not revenge-trade a loss, does not check its P&L at midnight and swear to quit. Those are things you do. The asymmetry is not just technological. It is behavioural, and it is permanent: the side of the market that feels nothing is taking money, in aggregate, from the side that feels everything.

Nobody had to cheat

The uncomfortable thing about the scoreboard is how little villainy it requires. No front-running scandal, no rigged prints, no vanished broker is needed to explain ₹61,000 crore a year. The structure does the work by itself: a negative-sum game after costs, a professional side running algorithms, a retail side recruited by the crore, and a toll on every order regardless of outcome. SEBI's study is not an exposé of rule-breaking. It is an audit of the rules working exactly as written. The house always wins not because the game is fixed, but because the game is the fix.

Why they always have a chair for you

Understand this and the whole industry's behaviour snaps into focus: the notifications, the referral bonuses, the expiry-day push alerts, the "insights" tab, the free trading courses inside the broker's own app. A market where the professionals collect ₹61,000 crore a year needs a constant supply of people to collect it from. You are not the customer of this market in any meaningful sense. You are its input. The machines are the customers - of the exchanges, of the brokers, of the data vendors - and you are what is being served.

The regulator counted it all and published it. The house always wins. Now, for the first time, you can read exactly how much it won, and from whom.