Here’s a question that quietly sorts every decision-maker into two camps: if you make a reckless bet and it pays off huge, was it a good decision?
Most of the world says yes, because most of the world reads the scoreboard and calls it the game. And that’s the direct answer to what process-versus-outcome thinking means: judge decisions by the quality of the process that produced them — the information, the reasoning, the execution — not by how they happened to land. Outcomes are process plus variance, and variance is enormous, unowned, and utterly indifferent to you. Grade yourself on the part you control, because grading yourself on the part you don’t will eventually teach you exactly the wrong lessons at exactly the wrong size.
Usual note for the Capital pillar: education and personal experience, not advice. What follows includes the single most embarrassing process failure of my life, which is precisely what qualifies it for the syllabus.
The scoreboard lies in both directions
Poker players have the cleanest name for the error — Annie Duke wrote a whole book circling it — they call it resulting: deciding a choice was good because it worked, or bad because it didn’t. Poker had to name it because poker makes the lie visible fast: you can play a hand perfectly and lose the pot, play it drunk and stack chips. Nobody who's serious about the game grades on pots. They grade on decisions, because decisions repeat and pots don’t.
Markets are poker with slower feedback and higher stakes, and life is markets with no dealer. The lottery winner did not make a good investment. The driver who arrives safely without a seatbelt did not have a good safety process. And — the direction people resist — the founder who did everything right and got flattened by timing didn’t necessarily decide badly. The scoreboard lies both ways. It flatters bad process with lucky wins and slanders good process with unlucky losses, and if you can’t decouple the two, every result makes you slightly dumber with slightly more confidence.
The email that never cashed
Now my receipt, because this piece would be theory without it. For years, I believed I had invested early in a company that went on to do very well. It was part of my internal highlight reel — evidence that my judgment was sharp early, a story I told myself when doubt showed up. The thesis had been right. I’d seen it before the crowd. I remembered deciding.
Then one day, digging through old email for something unrelated, I found the thread. The payment had never gone through. Some mundane failure at the last administrative step — and no follow-up from me, ever, because in my head the thing was done. The investment I’d been proud of for years did not exist. I hadn’t ridden the win or missed it. I had simply never verified the boring final link, and then built years of self-assessment on top of a transaction that wasn’t there.
Sit with how strange that is: I was right — genuinely, verifiably right about the company — and being right produced exactly nothing, because process is a chain and chains grade on their weakest link. Mine snapped at the least glamorous possible place: confirmation. The thinking was airtight. The wire was fiction.
Process includes the boring links
The lesson generalizes further than trading. When people hear “process,” they picture the intellectual part — the analysis, the thesis, the clever asymmetric read. But most real-world process failures aren’t intellectual. They’re operational: the order that didn’t fill, the document never countersigned, the follow-up never sent, the confirmation never checked. The unglamorous links, snapping quietly while all the attention guards the glamorous ones.
This is why serious professions run checklists — aviation and surgery didn’t adopt them because pilots and surgeons are dim, but because competence reliably fails at the mundane precisely because it’s mundane. Brilliance handles the hard parts and then walks past the easy ones with its hands in its pockets. So the fix is structural, not motivational: verification steps get written into the process as first-class citizens. “Confirm it actually happened” is a link in the chain, every time, with the same standing as the analysis — because my email thread can testify to what happens when it isn’t.
Why good process loses gracefully
Here’s the compounding argument for grading on process: repetition. Any single outcome is mostly noise — variance can swamp quality in either direction on any given day. But run enough reps and the noise cancels while the quality accumulates; good process converges toward good results the way a casino converges toward its edge, one indifferent hand at a time. You don’t need this trade to work. You need this kind of decision, made this way, to be profitable across a career of instances.
That reframe also changes how losing feels, which matters more than it sounds. A loss after sound process isn’t a verdict; it’s a toll — annoying, budgeted, survivable. The review question stops being “did it work?” and becomes “knowing only what I knew then, would I make the same call again?” If yes, the loss bought you nothing to fix and cost you nothing but money. If no, you found real tuition. Either way the stove taught something — but only the process-grader can tell which lesson was actually on offer.
The dangerous winners
And the inverse, which is where careers actually die: the bad-process win. It’s the most expensive event that can happen to a decision-maker, because it arrives disguised as validation. The reckless bet pays off, the skipped verification doesn’t bite this time, the oversized position works — and the lesson installed is that’s how it’s done, now with real money as the teacher’s aide. Every unearned win funds a bigger rerun of the same flaw, and the market lets the tab grow precisely as long as it takes to hurt.
The discipline is genuinely unnatural: audit your wins with the same suspicion as your losses. When the result was good and the process was garbage, log it as an accident you got paid for — an outcome to bank and a method to bury. The trader who survives isn’t the one who never gets lucky. It’s the one who refuses to let luck write the playbook.
Build the loop
The practical kit, descriptive as always: a decision journal — thesis, information at the time, expected range of outcomes — written before the result exists, because memory is a defense attorney. Verification built into every consequential move: money moved, order filled, document signed, confirmed by evidence rather than assumption. Reviews that grade the decision against the journal, not against the P&L. And one honest question at every postmortem, win or lose: which links held because they were strong, and which held because nothing happened to test them?
Outcomes are process plus variance. You only own one of those — so that’s the one you grade.
I was right about the company. I just never confirmed the one step that made being right mean anything. Right and rich are separated by process — verify the wire.
Tyk Sullivan
Serial entrepreneur in Tampa, Florida. Founder of Plutus Capital and NUMEN. More about Tyk →