Tick Forge audit report

Expert: Tick_Forge/TF_DemoEA. Source symbol: GOLD.pro, 2026.09.01 to 2026.10.27. Data: M1 bars.

Every symbol below is the broker's own history with only the date or the price level changed. The control is the unchanged history. Flags are rules of thumb, not statistics, and a flag is a question, not a verdict.

Flags

TFA_S14: the shifted data runs past today, so only the first part could be tested; compared on that overlap only (35 of 39 control entries).

TFA_S28: the shifted data runs past today, so only the first part could be tested; compared on that overlap only (25 of 39 control entries).

Results

Symbol	Change	Trades	Net	PF	Max equity DD	Entries reproduced
TFA_C	control (no change)	39	-2 046.00	0.31	24.57% (2 512.40)	100%
TFA_S7	shifted +7 days	39	-2 046.00	0.31	24.57% (2 512.40)	100%
TFA_S14	shifted +14 days	34	-1 982.70	0.28	22.88% (2 339.40)	97%
TFA_S28	shifted +28 days	24	-1 608.20	0.25	18.14% (1 854.10)	96%
TFA_L50	prices x0.5	39	-1 024.30	0.31	12.43% (1 257.40)	100%
TFA_L200	prices x2	39	-4 092.00	0.31	48.10% (5 024.80)	100%
TFA_MIR	mirrored moves	39	-2 243.70	0.32	27.13% (2 776.10)	100%

Entries reproduced = share of the control run's entries that reappear at the same moment after undoing the time shift. It is only meaningful for the calendar-shift rows.

TFA_S14: control entries that did not reappear (1)

2026.10.21 Wed

TFA_S28: control entries that did not reappear (1)

2026.10.07 Wed

What this does not show

A pass means the Expert did not depend on what was changed. It does not prove the Expert is safe or profitable. A failure points at something to investigate (dates, weekdays, price levels). Backtests on altered data are stress tests, not forecasts or trading results.

Generated by tf_audit.py (Tick Forge, MIT).