Track Record

How the model's past calls actually turned out. Kept honest on purpose.

Flip held 68% of days · Magnet touched 81% of days · 32 days tracked

68%

flip held · 21/31 days

81%

magnet touched · 25/31 days

days tracked

awaiting settlement

Daily record of gamma regime, flip level and magnet strikes against the session that followed.
Date	Regime	Flip	Spot @ snap	Net GEX	Magnet ↑	Magnet ↓	Day L — H	Close	Flip	Magnet	Status
SEP 222026	POS	770.19	773.55	$6.82B	775	772	772.57 — 775.14	773.38	HELD	ABOVE	Mostly matched
SEP 212026	POS	763.40	768.98	$8.97B	770	768	766.03 — 774.89	773.50	HELD	BOTH	Mostly matched
SEP 18

How this is scored. Each weekday morning the dashboard’s gamma flip level and its two biggest magnet strikes are recorded. After the close, that day’s high and low are pulled and judged: HELD means price never crossed to the other side of the flip level it started on; BROKE means it did. A magnet counts as touched if the session’s range reached that strike.

The status column. Two claims are made each morning — that price stays on its side of the flip level, and that the day’s range reaches a magnet strike. Both landing reads Mostly matched, one landing reads Partially matched, neither reads Did not match. Every settled day is listed in date order, misses included. Nothing here is filtered or sorted to put the good days on top.

Known bias. A daily bar carries no intraday timing, so the high and low include the part of the session before the snapshot was taken. That slightly over-counts both breaks and touches. It is a real limitation of free daily data, not a rounding detail.