Methodology
How phases are assigned, which rules are tested, and what the data says so far. Everything here is generated from the same research code that runs the backtests.
Universe
Every IPO in the Nasdaq IPO calendar's priced table, excluding SPACs and deals without a published offer price. Returns are measured from the final offer price, using split-adjusted daily closes. Day N means the Nth trading day after listing. Deals under $50M are tracked but reported separately, because they behave very differently (78% were below offer at day 120).
Lifecycle phases
- HypeFirst 20 trading days, or trading near its high with RSI above 60.
- DistributionMore than 15% below its running high, before the lock-up window.
- Lock-up windowFrom 20 trading days before to 10 after the estimated lock-up expiry.
- ValleyAfter the lock-up window, below its 50-day average and 30%+ off the high.
- AccumulationAfter the lock-up window, back above its 50-day average but below 90% of the high.
- RecoveryAfter the lock-up window, within 10% of its running high.
Phases use only data available on that day. Peaks and lows are never used as inputs.
What happened to IPOs of $50M or more
| Cohort | IPOs | Median peak | Peak day | Median drop from peak | Below offer at day 120 |
|---|---|---|---|---|---|
| 2024 | 64 | +36.7% | 40.5 | -49.2% | 55% n=64 |
| 2025 | 72 | +38.2% | 21 | -50.1% | 56% n=72 |
| 2026 | 89 | +40.8% | 20 | -44.1% | 56% n=43 |
IPOs tend to peak early and give back a large part of the gain. They do not reliably collapse below the offer price: about half are below it by day 120, and a fifth are up more than 50%. The spread between outcomes is the main fact, so every IPO page compares the deal with peers of the same size across day 1, day 20, day 60, day 120, day 180.
Frozen research rules (v1)
These rules were found and tested on every US operating-company IPO from 2014 to 2026, one held-out year at a time, then frozen. Signals dated from 2026-10-29 are recorded live in an append-only ledger and scored when their holding period ends. Edges are medians per signal against IPOs of the same year, size and age, in excess of the S&P 500 and net of 0.2% costs, from held-out tests on 2014–2026 IPOs run on Oct 28, 2026.
Every signal is published in the same impersonal form for every reader, for example “Rule O1: sell or short signal for 60 sessions”. Bullish rules publish buy signals. Bearish rules publish sell or short signals on deals of $50M or more and sell signals on deals under $50M. A signal states what a frozen rule found; it does not take any reader's situation into account. The rules run independently, so one IPO can carry buy and sell signals at the same time; each label reports only its own rule's history.
| Rule | Trigger | Published signal | Evidence |
|---|---|---|---|
| C1 · Capitulation | At least 60% below the highest close since listing, with RSI(14) below 30, after trading day 40 (first time only). | Buy signal for 40 sessions. | +3.1% edge; 13 of 13 years; 1,016 signals |
| E1 · Weak earnings reaction | After the first or second earnings release (SEC 8-K item 2.02), the stock trails the S&P 500 by 10% or more over the two sessions around the release. Deals of $50M or more. | Sell or short signal for 60 sessions. | +4.9% edge on $50M–$250M deals (84 signals) and +5.3% over $250M (108); 10 of 12 years; deals under $50M excluded (−3.0%) |
| O1 · Lock-up overhang | Twenty sessions before the unlock, locked shares are at least 6.75 times the shares sold in the IPO (top third of 2014–2025 deals). | Sell or short signal until 40 sessions after the unlock; sell signal for deals under $50M. | 386 flagged of 1,130 IPOs did 8.5 points worse than same-year, same-size IPOs; 10 of 13 years |
| F1 · Microcap deal | The IPO raised less than $50M. Checked at trading day 20. | Sell signal until day 120; shorting is usually impractical. | 10.4 points worse than larger deals; 11 of 13 years |
| T1 · Pre-unlock quality | Twenty sessions before the unlock: offer price above $11.20 and the close no more than 91% above the offer price. | Buy signal until 40 sessions after the unlock. | Beat other IPOs by a median 15.0 points in 12 of 13 held-out years |
| R20 · Day-20 risk tier | Risk score from eight price features at trading day 20, in the top third (bearish) or bottom third (bullish) of 2014–2025 scores. | Top third: sell or short signal (sell signal for deals under $50M). Bottom third: buy signal. Both until day 120. | 13.9-point spread between thirds; 11 of 13 years; walk-forward +8.3 |
| RL · Pre-unlock risk tier | The same kind of risk score, fitted 20 sessions before the unlock, in the top third (bearish) or bottom third (bullish) of 2014–2025 scores. | Top third: sell or short signal (sell signal for deals under $50M). Bottom third: buy signal. Both until 40 sessions after the unlock. | 11.6-point spread; 12 of 13 years; walk-forward +11.4 |
| U1 · Slide into the unlockborderline | Fell 30% or more from 20 sessions before the unlock to 2 sessions after it. | Sell or short signal for 40 sessions; sell signal for deals under $50M. | +5.7% edge; 6 of 8 years; interval touches zero |
The original momentum-entry and valley-entry ideas failed these tests and are not used. Further candidate rules are tracked privately until live results confirm them.
Do flagged stocks move the way the signal says?
The rules passed on returns relative to IPOs of the same year, size and age. Most IPOs fall after listing, so a buy signal can beat its peers and still lose money. The historical figures are absolute returns, without that adjustment.
| Rule | Signal | Sizes | Signals | Fell | Rose | Median return | Median vs S&P 500 | Rallied 25%+ | Dropped 25%+ |
|---|---|---|---|---|---|---|---|---|---|
C1 Capitulation | Buy | all deal sizes | 1,009 | 51% | 49% | -0.5% | -3.3% | 45% | 38% |
| C1 | Buy | deals of $50M and over | 652 | 47% | 53% | +2.1% | -1.1% | 43% | 33% |
| C1 | Buy | deals under $50M | 357 | 57% | 42% | -6.8% | -9.0% | 49% | 49% |
E1 Weak earnings reaction | Sell or short | deals of $50M and over | 186 | 70% | 30% | -13.6% | -15.5% | 33% | 51% |
O1 Lock-up overhang | Sell or short; sell under $50M | all deal sizes | 369 | 69% | 30% | -15.0% | -19.8% | 38% | 53% |
| O1 | Sell or short | deals of $50M and over | 191 | 69% | 31% | -13.4% | -17.2% | 28% | 47% |
| O1 | Sell | deals under $50M | 178 | 69% | 28% | -17.4% | -22.4% | 49% | 61% |
F1 Microcap deal | Sell | deals under $50M | 387 | 73% | 27% | -29.0% | -35.3% | 50% | 74% |
T1 Pre-unlock quality | Buy | all deal sizes | 673 | 52% | 48% | -1.7% | -4.5% | 32% | 34% |
| T1 | Buy | deals of $50M and over | 642 | 53% | 47% | -2.6% | -5.1% | 32% | 35% |
| T1 | Buy | deals under $50M | 31 | 35% | 65% | +5.1% | +1.6% | 42% | 13% |
R20 Day-20 risk tier | Sell or short; sell under $50M | all deal sizes | 398 | 75% | 25% | -29.9% | -37.0% | 52% | 76% |
| R20 | Sell or short | deals of $50M and over | 118 | 70% | 30% | -24.6% | -29.4% | 55% | 67% |
| R20 | Sell | deals under $50M | 280 | 76% | 23% | -33.9% | -39.3% | 50% | 79% |
R20 Day-20 risk tier | Buy | all deal sizes | 398 | 54% | 46% | -2.8% | -8.3% | 42% | 39% |
| R20 | Buy | deals of $50M and over | 389 | 54% | 46% | -2.7% | -8.2% | 42% | 39% |
| R20 | Buy | deals under $50M | 9* | 56% | 44% | -17.1% | -21.6% | 44% | 44% |
RL Pre-unlock risk tier | Sell or short; sell under $50M | all deal sizes | 388 | 69% | 30% | -16.0% | -19.5% | 43% | 60% |
| RL | Sell or short | deals of $50M and over | 155 | 70% | 30% | -18.5% | -19.9% | 34% | 61% |
| RL | Sell | deals under $50M | 233 | 69% | 30% | -13.9% | -18.4% | 49% | 58% |
RL Pre-unlock risk tier | Buy | all deal sizes | 388 | 46% | 53% | +0.9% | -2.0% | 27% | 23% |
| RL | Buy | deals of $50M and over | 329 | 46% | 54% | +1.3% | -1.7% | 26% | 22% |
| RL | Buy | deals under $50M | 59 | 47% | 49% | 0.0% | -3.5% | 32% | 27% |
U1borderline Slide into the unlock | Sell or short; sell under $50M | all deal sizes | 116 | 69% | 30% | -15.3% | -17.2% | 37% | 52% |
| U1 | Sell or short | deals of $50M and over | 60 | 67% | 33% | -11.4% | -12.5% | 32% | 43% |
| U1 | Sell | deals under $50M | 56 | 71% | 27% | -17.0% | -22.5% | 43% | 61% |
Frozen v1 rules on 2014–2025 IPOs; figures computed Oct 30, 2026.
* Fewer than 30 signals: too few to rely on. Signals in such a group show the all-sizes figures instead.
Past figures cover all 4,312 signals the published frozen v1 rules produce on 2014–2025 IPOs, close to close over each signal's holding period. Returns start at the close of the signal session; signals are published after that close, so the first price available to a reader is the next session's. Only 1,192 of 2,368 operating-company IPOs from those years have usable Yahoo Finance price data; most of the rest no longer trade under their IPO ticker (delisted, acquired or renamed), so the figures probably understate declines. These are hypothetical results, not trades: the rules were selected on these same years with the benefit of hindsight, so the figures likely overstate what live signals will do. The live track record is the real test. Trading costs and borrow fees are not included. Past results do not predict future results.
Short-selling risk: losses on a short position have no limit and can exceed the amount put up; brokers can demand more margin, recall borrowed shares or close the position; borrow fees and any dividends are owed while short and are not in the figures shown; recent IPOs are often hard to borrow and prone to squeezes.
Deal under $50M: published as a sell signal only, because shorting such shares is usually impractical (hard to borrow, high borrow fees, frequent squeezes).