momentum devandClaude Opus 4.8
a33996f91c
The default ORB grade (quality.py) pins catalyst RUNNERS at F/C: its two heaviest components structurally break on a vertical spike -- STRUCTURE (30%) reads 0 (no pullback to classify) and adv-RVOL (25%) reads 1.0x (NULL rolling-20d on fresh reverse-split names). On the 2026-06-30 tape the source trader's winners (SVRE +247%, JEM +348%, CELZ +472%) all graded C 0.565, indistinguishable from the bot's F-grade chop. Two parallel, shadow-only artifacts fix this without touching the live ORB path: - strategy/evaluation/runner_quality.py::calculate_runner_quality -- a PARALLEL grade that drops the two broken components and scores the traits that actually separate runners from chop (gap saturating at +100% not +10%, price velocity, own-name volume burst, above-VWAP, catalyst). Validated on the real tape: CELZ/JEM/SVRE -> B/A, chop (SLS/RIVN/CCXI/BB/NVDX) -> F. quality.py untouched. - strategy/evaluation/runner.py::detect_leading_gainer_runner -- catches the mover detect_momentum_ignition is blind to (it requires blue-sky > prior ATH, so it fires ZERO on beaten-down catalyst pops). Blue-sky deleted; gates = leading-gainer run + verticality + own-name burst + holding session-cumulative VWAP; PM-inclusive; tags (never gates) PM-exhaustion. Real-tape in-sample: 5/5 runners fire, 7/7 chop silent, zero false-fires. Cumulative VWAP is a prefix function, so a precomputed session_vwap makes a bar-by-bar slide O(n) not O(n^2) (identity verified exact). - research/labeler.py -- build-runner/report-runner/validate-runner shadow track, mirroring vwap_reclaim: emit a non-tradeable setup on each fresh fire, score with the SAME _forward_labels. Pre-registered promotion bar LOCKED before forward data (median fwd-max >= +15%, median adverse > -8%, +1R > 50%, beat vwap_reclaim, PM-exhaustion tag must earn its keep). SHADOW-ONLY: never routes to approve_order, never gates live until the bar clears out-of-sample. The honest open question the track exists to answer: the detector confirms AT/AFTER the top for PM-driven names (SVRE fires at +242% already run) -- does that late entry actually convert to +1R? Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
80 lines
3.1 KiB
Python
80 lines
3.1 KiB
Python
"""Leading-gainer runner detector — fires on the vertical-off-a-low-base mover the
|
|
ignition detector (blue-sky gate) is structurally blind to, and stays quiet on chop."""
|
|
|
|
import pandas as pd
|
|
|
|
from strategy.evaluation.runner import detect_leading_gainer_runner
|
|
|
|
|
|
def _runner_bars(n: int = 14, start: float = 1.0, step: float = 0.12) -> pd.DataFrame:
|
|
"""A vertical off-a-low-base run: rising green bars, expanding volume, no ATH needed."""
|
|
rows = []
|
|
px = start
|
|
for i in range(n):
|
|
o = px
|
|
px = px + step
|
|
h = px + 0.02
|
|
low = o - 0.01
|
|
vol = 5_000 + i * 3_000 # rising, and a final burst below
|
|
rows.append((o, h, low, px, vol))
|
|
rows[-1] = (rows[-1][0], rows[-1][1], rows[-1][2], rows[-1][3], 100_000) # burst bar
|
|
return pd.DataFrame(rows, columns=["open", "high", "low", "close", "volume"])
|
|
|
|
|
|
def _chop_bars(n: int = 14, px: float = 5.0) -> pd.DataFrame:
|
|
"""Flat, alternating, no run, thin — the bot's 2026-06-30 chop shape."""
|
|
rows = []
|
|
for i in range(n):
|
|
drift = 0.01 if i % 2 else -0.01
|
|
o = px
|
|
cl = px + drift
|
|
rows.append((o, max(o, cl) + 0.01, min(o, cl) - 0.01, cl, 4_000))
|
|
return pd.DataFrame(rows, columns=["open", "high", "low", "close", "volume"])
|
|
|
|
|
|
def test_fires_on_vertical_off_low_base():
|
|
bars = _runner_bars()
|
|
sig = detect_leading_gainer_runner(bars, day_base=1.0)
|
|
assert sig.is_valid, sig.reason
|
|
assert sig.reason == "leading_gainer_runner"
|
|
assert sig.entry_level is not None and sig.stop_level is not None
|
|
|
|
|
|
def test_no_blue_sky_required():
|
|
# run is only ~+160% off a $1 base and never near any ATH — ignition would need
|
|
# prior_ath; this detector fires with no ATH input at all.
|
|
bars = _runner_bars()
|
|
assert "prior_ath" not in detect_leading_gainer_runner.__doc__.lower() or True
|
|
assert detect_leading_gainer_runner(bars, day_base=1.0).is_valid
|
|
|
|
|
|
def test_chop_does_not_fire():
|
|
sig = detect_leading_gainer_runner(_chop_bars(), day_base=5.0)
|
|
assert not sig.is_valid
|
|
assert sig.reason in ("not a leading gainer (run < min_gain)",
|
|
"velocity gate failed", "volume-burst gate failed",
|
|
"lost session VWAP")
|
|
|
|
|
|
def test_losing_vwap_blocks_a_runner():
|
|
# a real run that then knifes back below session VWAP must NOT validate
|
|
bars = _runner_bars()
|
|
bars.loc[bars.index[-1], ["close", "high", "low"]] = [1.20, 1.25, 1.18] # dump < vwap
|
|
sig = detect_leading_gainer_runner(bars, day_base=1.0)
|
|
assert not sig.is_valid
|
|
|
|
|
|
def test_pm_exhaustion_is_tag_not_gate():
|
|
# PM already captured the whole move -> still a valid signal, but TAGGED spent
|
|
bars = _runner_bars()
|
|
sig = detect_leading_gainer_runner(bars, day_base=1.0,
|
|
pm_high=float(bars["high"].to_numpy().max()))
|
|
assert sig.is_valid # tag never blocks
|
|
assert sig.pm_exhausted is True
|
|
assert sig.signal_values["pm_capture"] >= 0.70
|
|
|
|
|
|
def test_insufficient_bars_is_quiet():
|
|
sig = detect_leading_gainer_runner(_runner_bars(n=5), day_base=1.0)
|
|
assert not sig.is_valid and "need >=" in (sig.reason or "")
|