The split

It breaks on the idea premium. Someone had the idea, so someone wants 60/40. The YC consensus, built from watching thousands of these age across a decade: equal or near-equal splits, four-year vesting with a one-year cliff, decided early. The reasoning isn't fairness for its own sake; it's that the idea was a rounding error over a ten-year game. Execution is the game. And an ungenerous split poisons year six from inside year one: a cofounder grinding through the struggle at 30% remembers the number every bad week, and bad weeks are most weeks.

It breaks worse without vesting. The schedule matters more than the ratio, because the catastrophic outcome isn't a slightly unfair split. It's a departed cofounder holding a quarter of the company forever: dead equity that poisons every future financing, mocks everyone still working, and cannot be clawed back by any amount of being right. The cliff and the schedule are what make any ratio survivable.

1. Run the coalition math before signing anything. On paper, before signatures: who can remove whom, alone or together, at each vesting milestone? Model it at month 6, at the cliff, at year two, at year four, with a board seat added, with an investor added. If one answer is "nobody can remove anyone, ever," you've built a deadlock machine. If another is "he can fire me the day after my cliff," you've built a trap and volunteered for it. This math takes an hour, almost nobody runs it, and it's the cheapest hour in the whole system.

2. Default to equal, then test the exceptions honestly. Start at equal and force deviations to justify themselves in writing. The honest exception is commitment asymmetry: full-time against part-time, capital in against not, joined at zero against joined at month nine. Those are real, priceable differences. Price them explicitly, in the document, never silently in resentment. A cofounder who accepts 50/50 while privately believing he deserves 65 hasn't agreed; he's deferred, and deferred grievances collect interest at a rate no cap table shows.

3. Vest everyone, founders included. Four years, one-year cliff, no exceptions, including you. Vesting isn't distrust; it protects every future version of the company from every departed version of a founder, and you are as capable of departing as anyone. Know your acceleration terms (single versus double trigger) and what they do in an acquisition, on purpose rather than by template.

4. Write the control terms, because equity alone doesn't decide. Board seats and how they change with financing. Vote thresholds for the calls that matter: selling, raising, firing an executive, firing a founder. Who's CEO, and, in writing, what it takes to change that. Two people holding 50/50 with no tiebreak haven't split control; they've deferred a fight to the worst possible moment, and the worst possible moment is precisely when it will arrive.

5. Set the rules while everyone still likes each other. From the documented wreckage of cofounder wars: once things go wrong, confrontations ARE governance moves, whether intended or not, and coalitions form within days of the first hard conversation. Rules written in the honeymoon execute themselves in the divorce. Rules written during the divorce are war.

• The idea premium. Rounding error. Say it out loud until it stops stinging.
• Handshake splits. Unwritten means renegotiable by whoever gains leverage, and leverage always moves.
• Skipping the cliff out of trust. The cliff exists because month eleven knows things month one couldn't.
• Part-time at parity. Asymmetry unpriced becomes resentment financed.
• 50/50 with no tiebreak and no removal path. Deadlock is a failure mode, not fairness.
• Reopening a signed split without new commitment facts. That's leverage talking, and it should be named as such, in the room, when it happens.

Two technical founders, three months in, revenue zero. A wants 60/40: the idea was his, and he started six weeks earlier. B has been full-time since day one; A is still at his job until the seed closes. The reframe runs: rights, not credit. The idea premium prices at what the record says it's worth, which is nothing. But the commitment asymmetry is real, and it runs the other way: B is full-time and A isn't yet. So it gets priced explicitly: 52/48 in B's favor until A goes full-time, converting to 50/50 the day he does, written into the vesting schedule instead of carried as a grudge. The coalition math finds the deadlock (50/50, two board seats, no tiebreak) and fixes it in the honeymoon: A is CEO, B leads product, a mutually agreed third seat arrives with the seed, and firing a founder requires board majority. Four-year vesting, one-year cliffs, acceleration terms chosen on purpose. The whole package: two evenings and $2k of lawyer time.

At month nineteen they hit their worst stretch, and discover the 2am argument has nowhere catastrophic to go. The rules were already written, by two people who liked each other. Which is exactly who you want writing them.