IPO / COHORT ANALYSIS

Compare IPOs without trusting the average

A normalized chart can reveal useful questions about early trading. It can also make a noisy collection of very different businesses look more predictable than it is.

JBOT editorial · Published September 20, 2026 · Updated September 21, 2026 · Educational guide

Choose the starting point explicitly

An IPO's offering price, first exchange trade, day-one open and day-one close are different possible baselines. A public-market buyer may not have had access to the offering price. For an apples-to-apples chart, choose one baseline and state it clearly.

Index observations by trading session rather than calendar date. Missing bars, holidays, halts and changed tickers need deliberate handling. A blank observation should not silently become a zero return.

Look beyond the average line

• Compare the mean with the median. A few extreme winners can pull the mean away from the typical experience.
• Inspect the sample count at each age. Fewer observations near the end of a chart can change its interpretation.
• Separate cohorts by relevant characteristics where sample size allows: period, sector, company size or liquidity.
• Include unsuccessful and delisted names where reliable data is available. A survivors-only sample can flatter a strategy.

Read the company documents too

The prospectus can explain the business, risks, use of proceeds and share structure. Lockup arrangements and other restrictions differ by issuer; do not infer an exact event date from a generic calendar rule.

Treat simulation bands as conditional scenarios

A Monte Carlo engine generates paths under selected assumptions. More paths can reduce simulation noise without correcting an unrealistic return model, a biased training set or missing information. One million paths do not imply one million independent pieces of evidence.

Judge the model on later, untouched observations: whether its intervals cover outcomes as expected, whether its errors change across cohorts, and whether any proposed decision rule survives realistic trading costs. Keep an untouched benchmark and compare against it before treating added complexity as progress.

General education, not personalized investment advice. IPOs can be volatile and illiquid. Simulation bands are not guaranteed price ranges or a promise of returns.