The empty seat.
Why Tally exists, in the plainest terms we can put it.
Traders don't lose because their ideas are bad. They lose because they don't follow their own rules — and AI made that worse, because now you can generate a brilliant-sounding reason to break them in seconds.
Every trading desk staffs a risk officer. Solo traders don't have one. Tally is that seat: rules written before money moves, frozen so you can't quietly edit them, enforced by something that isn't you at the moment you most want to negotiate.
You stop fighting yourself. Win or lose, you can say “I followed my system” — and prove it, because every closed trade mints a receipt.
Every desk staffs two seats.
The analyst finds the idea, builds the case, wants to put on risk. The risk officer never finds ideas — they set the mandate before the trade and enforce the exit when it's breached. Critically, the risk officer is not the trader's adversary; they're the trader's protection against their own conviction.
The whole structure exists because of one hard-won institutional truth: the person holding the conviction cannot be the person who decides when the conviction has failed. They're too invested — and they're the most invested at exactly the moment the decision matters.
Solo traders hold both seats. And the analyst always wins the argument, because the analyst is the one who's excited, and who's present. That's where the losses come from. Not bad ideas — an unstaffed seat.
Then AI made the analyst seat superhuman. Research that took a team takes a prompt now. But it made the rationalizer superhuman too — the same model that found your edge will happily produce a compelling case for abandoning the exit rule you wrote, if you ask it the right way, at 2am, while the position is red. AI upgraded one seat to genius level and left the other one empty. The imbalance got worse, not better.
Tally sits in the empty seat.
The first receipt is ours.
Tally's founding trade was a loss: a yen-carry thesis that didn't transmit. It exited by the written rules with about 78% of the premium preserved, and it's graded in public — Process B, Outcome D. A disciplined loss, honestly recorded, beats a lucky win. That's the whole standard, and we hold ourselves to it first.