Turn any decision into
a reproducible investment report.
Time, money, labour, assets, opportunity cost — every commitment is an investment. Decision Lab computes your win rate, payoff ratio, expected value, Kelly position size and a Go / No-Go recommendation, then ranks all your options by score = p × b.
✓ 5 minutes ✓ No account ✓ First 5 report pages free ✓ Formulas and Python code fully open
10
ready-made templates
12
page feasibility report
10,000
Monte Carlo runs / report
From the Decision Lab business plan
“We turn every life decision into a quantifiable investment thesis.”
3 steps · 5 minutes
From gut feel to investment-grade analysis — one form in between
No SaaS learning curve, no spreadsheet anxiety, no LLM hallucination risk. Every number ships with its formula and its Python/TypeScript implementation.
1. Describe what you're committing
Tell the wizard how much cash, time, headcount, physical assets and opportunity cost you're prepared to put in. Ten one-click templates give you sensible starting values.
2. The platform does the maths
Expected value, Kelly position size, RAROC, 10,000 Monte Carlo runs, a six-variable tornado, three-scenario analysis and a counterfactual matrix — computed in under a second, every formula inspectable.
3. Get a 12-page report
Read it online and download the first 5 pages as PDF free. The GO / WATCH / GO_REDUCED / NO_GO recommendation is tuned to your profile — student, founder, family decision-maker and so on.
Live example
Real numbers on which decisions are worth making
These six opportunities come from the platform's seed templates, ranked by Decision Lab using one consistent formula. Open any row for the full 12-page report — no login, no payment.
- Ranking:
score = win rate × payoff ratio, tie-broken by RAROC then EV. - Colour: green = GO, gold = GO_REDUCED, blue = WATCH, red = NO_GO.
- Bubble size = equivalent capital committed.
Opportunity quadrant
| # | Opportunity | Win rate | Payoff | EV | Action |
|---|---|---|---|---|---|
| 1 | Professional certification (PMP) | 60% | 5.7× | $17,113 | GO_REDUCED · Proceed at reduced size |
| 2 | Index-fund dollar-cost averaging over 5 years | 77% | 2.8× | $117,496 | GO_REDUCED · Proceed at reduced size |
| 3 | Big Tech to late-stage startup | 35% | 2.9× | $38,514 | GO · Proceed |
| 4 | Buying a first home | 64% | 1.2× | $23,925 | GO_REDUCED · Proceed at reduced size |
| 5 | Overseas STEM degree | 23% | 2.2× | -$19,616 | NO_GO · Do not proceed |
| 6 | Bootstrapped micro-SaaS | 13% | 3.5× | -$7,791 | NO_GO · Do not proceed |
Show the maths
No black box: four formulas carry the whole report
Every number Decision Lab produces comes from four well-established formulas plus one explicit decision tree. Each report ships with:
- · a Python reference implementation (identical to
models/feasibility_model.py) - · a TypeScript implementation (identical to
lib/feasibility/) - · 10,000 Monte Carlo runs producing the real P5/P50/P95 distribution
- · ±20% tornado sensitivity analysis, naming the two variables worth researching further
Expected value (EV)
EV = capital × ( p × b − (1 − p) )capital: the equivalent capital of your five resource types · p: win rate · b: payoff ratio
Kelly optimal fraction f*
f* = ( b · p − (1 − p) ) / b (clipped at 0)Gives a mathematical answer to "how much of my risk budget should this take?"
Risk-adjusted return (RAROC)
RAROC = EV / ( capital × σ )The same risk score used by hedge funds, investment banks and the Basel framework.
Recommended action
Action = go_no_go(EV, f*, risk_budget, RAROC)
ev ≤ 0 → NO_GO
raroc < floor or kelly < f0 → WATCH
kelly ≤ risk_budget → GO
else → GO_REDUCEDTranslates the metrics into an action using the risk budget from your profile.
Why Decision Lab
Seventy years of settled mathematics, applied to ordinary decisions
We invent no new theory. We take the frameworks Kelly, Markowitz, Sharpe and Kahneman & Tversky have validated over seventy years and turn them into a report a ten-year-old can read.
34.8M
US small businesses (SBA Office of Advocacy, 2024)
20.4%
of new US establishments close within their first year (BLS Business Employment Dynamics)
65.3%
of new US establishments close within ten years (BLS Business Employment Dynamics)
2×
loss-aversion coefficient — losses hurt about twice as much as equivalent gains (Kahneman & Tversky, 1979)
Methodological foundations · four cornerstones
- 1956
Kelly Criterion
John L. Kelly Jr. · Bell System Technical Journal
Used for: optimal position size for a single opportunity
- 1952
Markowitz Portfolio
Harry Markowitz · The Journal of Finance
Used for: allocating resources across several opportunities
- 1966
Sharpe / RAROC
William Sharpe · The Journal of Business
Used for: ranking by risk-adjusted return
- 1979
Prospect Theory
Kahneman & Tversky · Econometrica
Used for: the loss-aversion coefficient in user profiles
Full data sources and methodology are in chapter 5 of the Decision Lab business plan.
FAQ
7 questions about Decision Lab
What is Decision Lab?
Decision Lab is an investment-feasibility platform. It treats any commitment — money, time, labour, physical assets, opportunity cost — as an investment, then uses win rate, payoff ratio, expected value, Kelly position sizing, RAROC and Monte Carlo simulation to produce a Go / No-Go recommendation and a reproducible 12-page feasibility report.
Do I need an account, and does it cost anything?
No account is required; the platform works the moment you open it. Generating a report is free and the first 5 pages of the PDF download free. The full 12-page PDF unlocks with an email address, also free.
How is the report calculated?
It rests on four classical frameworks: the Kelly criterion (1956) for optimal position size; expected value EV = win rate × gain − (1 − win rate) × loss; RAROC for risk-adjusted ranking; and a 10,000-run Monte Carlo simulation for the outcome distribution, with a tornado chart for single-factor sensitivity. Every formula is shown in the report.
What kinds of decision can it evaluate?
Ten templates ship with the platform: a side-project SaaS, graduate school, joining a late-stage startup, buying a home, studying abroad in a STEM field, a food-and-beverage franchise, building a creator business, a professional certification (CFA/PMP), opening a coffee shop, and index-fund investing. You can also start from scratch and enter the resources and assumptions for any decision of your own.
Can the results be reproduced and checked?
Yes. The engine has two independent implementations, TypeScript and Python, kept in agreement by golden-parity tests. Template defaults, industry benchmarks and formulas are all open, and the code is on GitHub for anyone to audit and re-run.
Where do the win-rate benchmarks come from?
The platform maintains Beta(α, β) win-rate priors for eight decision categories. The English-market priors are drawn from public US sources — BLS Business Employment Dynamics establishment survival, BLS and NCES education and earnings data, NAR and Case-Shiller housing data, and long-run S&P 500 total returns — and each is listed with its source in research/sources-en.csv. Priors are updated with Bayesian posteriors as users log real outcomes.
Is my data safe?
The platform asks for no account and no personally identifying information; visitors are distinguished by an anonymous session cookie. The decision data you enter is used only to generate your own report, and a report is shared only if you deliberately create a share link.
Five minutes from now you'll have a report you'll regret less.
No sign-up · Every formula auditable · First 5 report pages free