Screen and Size

Phase 2 · deployable results

Screen and Size

What survived the full verification stack: one axis that decides what is eligible to hold and what to avoid, and two ways — vendor and price-only — to decide how large the position should be. Neither forecasts returns for an individual name — that capability was tested for extensively and certified absent.

Russell 2000 · S&P 500 26-week horizon One-touch test 2023+ Verified through 2026-09-15
The job — what is eligible to hold, and what to avoid

The drawdown axis

hi252 >= -0.12269   — keep
hi252 <= -0.4947   — avoid
Placebo-clean Median-confirmed Positive on 3 PIT universes Monotone across the range One axis, two uses — not two rules

The whole effect in one table

Distance below the one-year high orders both outcomes monotonically. Everything else in this section is a way of cutting this ladder.

All 342k stock-weeks in the test window, bucketed by distance below the 252-day high. Bottom-quintile rate against a 0.1998 base; mean 26-week return relative to the index.
Distance below one-year highRowsBottom-quintile rateMean relative
Within 5%49,0990.116−320bp
5–12% below48,1440.113−334bp
11.5–22% below67,8250.140−496bp
22–33% below53,1420.194−734bp
33–45% below45,2220.256−971bp
More than 45% below78,9710.323−1,137bp

The odds of landing in the worst return quintile nearly triple from the top bucket to the bottom. The first two rows are flat, which is what makes the ~12% cut below a sensible place to draw a line rather than an arbitrary one: the effect is flat inside 12% and degrades steadily outside it.

What it measures

hi252 is the log distance below the trailing 252-day high: log(P_today / max(P over last 252 trading days)). It is always zero or negative. The threshold −0.12269 puts the price at 88.5% of its one-year high or better — in plain terms, within about 11.5% of the one-year high.

Use one — keep only the top of the ladder

  1. Build your candidate list however you normally do.
  2. Compute hi252 weekly for every candidate.
  3. Drop any name sitting more than ~11.5% below its one-year high.

It is an exclusion applied to a list you were already building — never a signal to buy on its own. Refresh weekly; the evidence horizon is 26 weeks.

Use two — avoid the bottom of the ladder

The same axis read from the other end. Names more than ~39% below their one-year high (hi252 <= -0.4947) hit the worst return quintile at 0.3119 against a 0.1998 base — a 1.6× enrichment across 100,282 stock-weeks and 1,752 names, at time-z 9.94. This is the strongest time-axis evidence anywhere in the project.

Avoidance is the cheapest thing to act on: declining to buy costs nothing, so there is no transaction-cost hurdle to clear. Survivorship bias also works in its favour — the delisted names missing from the data sit disproportionately in this bucket, so the true enrichment is understated rather than flattered.

A leg we removed

This screen was originally banked with a momentum leg attached (mom12_1 >= 0.1764 & hi252 <= -0.4947). Ablation killed it. Momentum alone selects names that are less likely to land in the worst quintile (0.1812 against the 0.1998 base, time-z −2.56) — it is protective, not a risk marker. Adding it lifts precision slightly, from 0.3119 to 0.3282, while costing ten times the coverage and halving the time-clustered evidence, from 9.94 to 4.85. Use the drawdown leg alone.

What it bought

Within high-momentum names, 26-week returns relative to the index. Test window 2023+.
SetRowsMeanMedian
Kept — near the high53,505−264.1bp+24.4bp
Excluded — deep drawdown53,506−751.7bp−546.9bp
Gap—+487.6bp+571.4bp

The median gap exceeds the mean gap, and the kept set's median is positive. That is the signature of a distribution shift rather than a trimmed tail — the usual objection, that excluding volatile names mechanically lifts a skewed mean, does not apply here. A random exclusion of the same size is inert: +189.5bp at time-z 1.73, which is simply the momentum baseline.

Tuning

The cut is a continuous tilt. Significance is flat across the whole range, so choose the cut that matches how many positions you want — not to maximize a statistic.
CutGain vs universeTime-zRows kept
Within 5% of high+482.7bp3.7529,039
Within 11.5% (banked)+436.6bp3.4853,505
Within 20%+393.1bp3.7170,577
Within 30%+336.6bp3.7884,048
Within 50%+254.9bp2.8696,842

Point-in-time universes, rebuilt from archived index holdings: +690bp (2022), +96bp (2024), +384bp (2025). Positive in every one, weakest in 2024.

Where it works

The ladder holds across the whole liquidity range, which matters for capacity — you do not have to trade thinly-traded names to get the effect.

Spread between the worst and best drawdown bucket, within Russell 2000 liquidity terciles.
Liquidity tercileBest bucketWorst bucketSpread
Least liquid third0.1180.333+0.215
Middle third0.1100.319+0.209
Most liquid third0.1200.309+0.189

The gradient itself is universe-invariant. On the same 2023+ window the S&P 500 ladder runs 0.143 → 0.344 (spread +0.202) against the Russell's 0.116 → 0.323 (+0.207) — the same per-name effect in both. A point-in-time 2022 membership audit confirms survivorship only flattens the S&P read slightly (+0.243 PIT versus +0.220 current). What differs is population: about 23% of Russell rows sit more than 45% below their high, against 2.5% in the S&P — the screen works everywhere, but the Russell offers roughly eight times as many names in the bucket worth avoiding.

An earlier version of this page attributed the flat S&P ladder to index quality-filtering. That comparison mixed date windows; matched windows show the invariance above.

The job — how large the position should be

Volatility-character forecast

peglt <= 0.57 & opinc_ttm <= 756.9
Both legs load-bearing Blind-recovered to 4 decimals Re-searched over full space Forecasts risk, not return

What it measures

PEG-long-term in the cheapest ~26% of names and trailing-twelve-month operating income in the smallest ~14% — under roughly $757M. Small companies that look cheap relative to their growth expectations.

It predicts that the stock will land in the top tercile of 63-day realized volatility, measured 26 weeks later: 0.836 against a 0.310 base rate, a 2.7× lift, from fundamentals alone with the outcome measured on independent price data.

How to use it

  1. Sizing. Flagged names deserve smaller positions for the same risk budget.
  2. Screening for stability. If the mandate needs low drawdown, exclude flagged names outright.
  3. Expectation setting. A flagged holding is very likely to swing hard within six months. Do not mistake that for the thesis breaking.

Why both legs stay

Test window, predicting top-tercile volatility. Base rate 0.3095.
RulePrecisionSignalsNamesTime-z
peglt leg alone0.424618,26622510.57
Size leg alone0.494610,31912716.99
Both together0.83571,2544021.86

Neither leg alone reaches even half the conjunction's precision. This is a genuine interaction — unlike several other multi-leg rules in this project, where ablation showed a leg was decoration. Do not drop either one.

Tuning

The size cut trades precision against coverage smoothly. The specific $757M is arbitrary; the leg itself is not removable, only re-tunable.
Operating income capPrecisionSignalsNames
$400M0.864377424
$757M (banked)0.83571,25440
$1,337M0.72892,77871
$3,000M0.61736,373119
$10,000M0.513912,933183
The job — sizing when you have no vendor data

Persistence map

predict 26wk-ahead tercile = current tercile   (vol: 126d estimator · liquidity: 63d · correlation: 126d)
Free data — prices only Time-z 24–117 over a decade Survivorship floor computed Forecasts risk, not return

What it does

Everything the volatility-character forecast does, at lower precision, from daily prices alone: a name's cross-sectional tercile of volatility, liquidity, and index correlation predicts its tercile 26 weeks out. No licensed columns, no vendor dependency.

Row-level accuracy vs a 1/3 base, 2016–2026, Russell 2000. Year-clustered z is the conservative axis.
AttributeAccuracyYear-zSurvivorship floor
Liquidity (63d $-volume)0.82784.9~0.72
Volatility (126d estimator)0.67727.7~0.57
Index correlation (126d)0.58024.0—

The 126-day volatility estimator is a verified upgrade over 63 days (+2.0pp on the one-touch test, positive in all six test half-years); 252 days over-lags. The floor column is the worst case if every delisted name absent from the data (~21% of historical membership) predicted at chance.

Use the vendor rule when you hold the license — its conjunction reaches 0.836 on a targeted slice. Use this map for everything else: it covers every name with a year of price history, and its liquidity forecast is the strongest-evidenced number in the project.

Replay: gp_persist r2k _RUT  ·  variant search: gp_persist improve r2k _RUT

Using both

One weekly cycle

The two rules compose naturally because they answer different questions. The drawdown axis decides what is eligible; the forecast decides how much.

  1. Build the candidate list.
  2. Drop names more than ~11.5% below their one-year high — or, if you only want the cheap version, drop the ones more than ~39% below and keep everything else.
  3. Run the volatility flags on the survivors.
  4. Size inversely to the forecast volatility character.

The two ways of cutting the ladder are not additive — they are the same axis. Pick the one that matches how much of the universe you are willing to give up.

Replaying each rule against the banked datasets:

gp_enum eval ytr_test.tsv ytr_test_ids.csv \
  "mom12_1 >= 0.1764 & hi252 >= -0.12269" rel26

gp_enum eval ytr_test.tsv ytr_test_ids.csv \
  "hi252 <= -0.4947"

gp_enum eval ckrisk_y_vol_test.tsv ckrisk_y_vol_test_ids.csv \
  "peglt <= 0.57 & opinc_ttm <= 756.9"

Limits to keep in front of you