What we do
Three mandates, one discipline.
Every hard decision has the same anatomy: a judgment, a cost of being wrong, and a process that was never measured. We work on all three.
-
Venture studio
Our own instruments, built and run end to end. The first is live. The next ones are pointed at gates where a bad yes is expensive and nobody has measured how inconsistent the yeses are.
-
Applied decision science
The same discipline on a client’s decisions. Where judgment enters the process, how much it varies, what the variation costs, and what an instrument would have to do to fix it.
-
Thinking in public
We write down how we reason about numbers and we publish it. The rubrics stay private. The reasoning behind them doesn’t.
The philosophy
Three disciplines, one question.
How do you tell whether a judgment is any good? Psychology, economics and computer science each answer a third of it.
-
01
Psychology
How people actually decide. Two experts given the same file reach different answers, and the same expert on a different day does too. Kahneman, Sibony and Sunstein called it noise. It is larger than anyone assumes until someone measures it.
-
02
Economics
What a wrong decision costs, and to whom. Base rates, expected value, the asymmetry between a bad yes and a missed no. The arithmetic decides which errors matter, and it rarely agrees with intuition.
-
03
Computer science and AI
How to build the instrument. A rubric written before the evidence, an anchor on every score, uncertainty that is computed rather than felt, and a model that scores case ninety exactly the way it scored case three.
You do not fix inconsistent judgment by training the judge. You fix it by changing the procedure the judgment happens inside. That is what an instrument is. The founder trained in all three.
Venture studio
The instruments.
One engine. A rubric written before the evidence. Factors scored one at a time, the verdict held until last. An anchor on every score, uncertainty declared, output comparable across cases. Pointed at gates where a bad yes is expensive.
-
Live
first-pass
Screens business ideas before anyone spends a quarter on them. Seven hard gates, twelve scored factors, three buyer profiles for accelerators, funds and operators, and a verdict that can say “there is no answer you could give me that changes this” and mean it arithmetically.
Built to kill about 80% of ideas in under a minute, for cents apiece.
-
Pipeline
Admissions
Selective admissions is a room of experienced readers disagreeing about the same file, and the disagreement is invisible because nobody scores the readers. The same engine, with a rubric written for the committee rather than the applicant.
-
Pipeline
Underwriting
Credit and insurance decisions where two underwriters price the same risk differently and the gap between them is a cost nobody has put on the books. Consistency is the product.
Same engine. New rubric per gate.
Method
How we think about numbers.
The rubrics are private. The reasoning behind them is not. Kahneman, Sibony and Sunstein call the remedy for noisy judgment decision hygiene. These six rules are our version of it.
-
01
Write the rubric before you see the evidence.
Adjust it mid-analysis and you fit conclusions to findings, and you lose the ability to compare anything to anything. The sciences call this pre-registration. It works here for the same reason.
-
02
Every score gets a written anchor.
A 0 to 5 with no anchor is astrology with decimals.
-
03
Declare what you don’t know, then compute what it could change.
A verdict that cannot say “no answer changes this” is not a verdict yet. It is a guess with a number on it.
-
04
Show the arithmetic on every number.
Including the one at the top of this page. A test that’s 99% accurate, looking for something that happens 0.1% of the time, is wrong about 90 times out of 91. That is not a flaw in the test. It is what the base rate does to it.
-
05
A process that greenlights everything is a random number generator.
The kill rate is the product. A screen that never says no has measured nothing.
-
06
Test the instrument the way you would test the judge.
Score the same case written three ways, facts held constant and only the prose changed, and watch how far the number moves. If style moves it, the rubric is scoring writing rather than substance.