Finance Isn’t Math. It’s Judgment.
The myth: finance is formulas and precision.
The reality: judgment under uncertainty.
The people who talk about their WACC to two decimal places or present a 12-tab forecast model with“exact” numbers three years out are telling you one thing: they have no clue what they’re doing. Precision in finance is theater. The math matters, but only in context. Numbers alone don’t make decisions. People do. The game is about making the best possible decision with the best available incomplete information and then adapting when the world refuses to match your spreadsheet. And it never does.
In practice, the best operators treat spreadsheets as instruments, not oracles. They know when to trust them, when to override them, and how to use them to align capital with intent. That’s the skill too few analysts, founders, or CFOs ever develop, and the one AI can’t replicate, at least not yet. While finance relies on quantitative tools, the real discipline is judgment under uncertainty.
Precision Is a Joke
I’ve watched analysts debate whether the WACC was 9.31% or 9.41%, as if that difference means anything in the real world. It doesn’t. Your cost of capital isn’t a fixed number handed down from above; it’s an approximation built on assumptions about risk, leverage, market conditions, and investor expectations, all of which can change before you finish the meeting and evolve over time.
Same story with forecasts and valuations. A five-year forecast is, at best, a directional guide to help you frame scenarios and tradeoffs. The numbers are not reality. They’re structured fiction about a future state that will never happen exactly as written. If someone tells you they can forecast with precision, they’re selling certainty that doesn’t exist.
The Real Game: Decision-Making Under Uncertainty
Real operators accept uncertainty and build their decision-making process around it. They run scenarios, not just “base, high, low.” They think in probabilities and bell-shaped distribution curves, not single outcomes. They map decision trees, assign likelihoods, and look for asymmetric payoffs where the upside dwarfs the downside.
And they account for option value. Not in the Options-pricing textbook sense of Black-Scholes, but in the strategic sense of keeping paths open. In most companies, this isn’t even in the spreadsheet, yet it’s often the difference between surviving and winning when conditions change.
Sometimes the best capital allocation decision isn’t the one with the highest modeled NPV, IRR, or MOIC, it’s the one that buys you flexibility to pivot when you see how the market evolves. Don’t forget the value of new information. Like holding off on an expansion plan until you see if your competitor’s supply chain collapse is temporary or permanent. Strategic optionality is invisible in most spreadsheets, but it’s worth a lot when the environment changes faster than your planning cycle.
AI Will Do the Math. You Still Have to Think.
AI can now run analyses in seconds or minutes that used to take analysts hours. It can pull comps, build a model, spit out a DCF, or benchmark industry metrics in minutes. That can be useful. It’s leverage.
But AI won’t tell you if the data is garbage. It won’t know that the “peer” it found has a different revenue model and GTM strategy, or that the market trading multiple it pulled was inflated by a one-off acquisition. The deal added debt that increased Enterprise Value but LTM EBITDA wasn’t adjusted to pro forma for the acquisition. You’d be surprised how often that one slips past “experts.”It won’t challenge your assumptions, ask whether your cost of capital is reasonably estimated, or remind you that the high case you’re excited about still leaves you exposed if the macro environment turns or market structure changes.
AI is a tool, not a substitute for judgment. The output’s only as good as the lens you’re looking through. And that lens comes from experience, context, and the ability to connect the numbers to reality. That’s why Forma teaches decision-making under uncertainty, not prompt hacks or spreadsheet acrobatics.
How to Sharpen Judgment in Finance
Judgment isn’t taught in most business schools, and you can’t download it from ChatGPT. You build it by combining structured analysis with real-world feedback loops. That means:
- Interrogate assumptions – Where did the input come from? How could it be wrong? What would make it change?
- Run scenarios and sensitivities – They’re different concepts. Don’t stop at best/worst/base. That’s lazy modeling. Understand the edges of the distribution. And when you run sensitivities, model the way real drivers move together in correlated fashion, not in isolation.
- Consider optionality – What’s the value of keeping this decision reversible for six months while new information comes in?
- Balance math with context – Does the result make sense in the market you operate in?
- Close the loop – Compare your forecasted outcomes to reality, learn where you were wrong, and feed that back into the next decision.
The Bottom Line
Finance is not about perfect answers. It’s about better decisions. Precision is for engineers; capital allocation is for decision-makers. Your job isn’t to produce a perfect model; it’s to use the model, the market, and your own judgment to make a choice that moves the business forward under uncertainty.
AI will get faster and smarter. But the best operators will still be the ones who know when the math is lying, when the base case is a trap, and when the “wrong” number on paper is the right call. That’s the part the spreadsheet never tells you. No amount of formatting hides that.