A profit factor of 999 can mean zero trades, not a hot streak

CL, ticker for Colgate-Palmolive, shows a profit factor of 999 under 52-week low reversion. Behind that number are 8 real trades, every single one a winner, a 100 percent win rate and a 67.92 percent return. NVDA shows a profit factor of 999 too, under NR7 volatility breakout. Behind that number are zero trades. Not zero losing trades. Zero trades of any kind, a return of 0.0 percent, a win rate of 0.0 percent, every other figure on the row reading exactly zero. The same number, 999, describes an 8 for 8 winning streak in one case and nothing happening at all in the other.

Profit factor is gross profit divided by gross loss. When gross loss is zero, that division has no real answer, so the underlying system substitutes 999 rather than leaving the field blank or writing infinity. That is a reasonable engineering choice. It is also a number that looks identical whether zero losses sat under eight genuine winning trades or under no trades whatsoever, and nothing about the figure itself signals which one produced it.

How common the cap is, and where the empty rows cluster

Across every sector's top-five and flop-five ticker list in all 22 published studies, 2,386 completed per-ticker results carry a profit factor figure. 150 of them, about 1 in 16, read exactly 999. Of those 150, 41 have zero total trades, the same empty pattern as NVDA above. Those 41 are not spread evenly. 36 belong to NR7 volatility breakout alone, and the other 5 belong to 52-week high momentum. Tenachine's guide on NR7's risk to reward ratio already found NR7 the one study on this site with an aggregate ratio below breakeven. A high count of zero-trade tickers sitting inside NR7's own top and flop lists is a second, separate signal pointing at the same study: a lot of NR7's 220-ticker universe produced little or nothing to measure.

What a zero-trade row still means to show up at all

A ticker with zero trades still carries `"success": true` in the underlying data and still shows up in a sector's top or flop five, sorted alongside tickers that actually traded. That is worth sitting with. Success here means the backtest ran without error, not that the strategy found anything to do. NVDA sitting in NR7's Information Technology top five with 0 trades is not NVDA underperforming or outperforming anything, it is NR7's entry rule never triggering for NVDA once across the ten-year window. A reader scanning that sector's top five for the best NR7 result on a familiar ticker would see NVDA's name and a 999 profit factor before noticing the 0 next to total trades.

The 109 remaining tickers with a genuine 999, the ones with 1 or more real trades, are not all suspect either. CL's 8 trades and CAT's 6, both under 52-week low reversion with 100 percent win rates and triple-digit returns, are real results worth taking seriously, precisely because a zero-loss streak over 5 or more trades is a much stronger claim than a zero-loss streak over 1 or 2. The cap is not evidence against any of these tickers. It only stops working as a ranking tool the moment two rows show the same 999 for entirely different reasons.

This is not a reason to treat NR7 volatility breakout, 52-week low reversion, or any ticker named here as broken or as a signal to trade. A backtest is a description of one rule applied to one universe over one historical window, and a zero-trade result says only that the rule and that ticker's price action never lined up in this window, not that the rule failed. Before reading any profit factor at or near 999 as a finding, check the trade count sitting next to it. A number capped at an arbitrary ceiling cannot tell a real streak from an empty row on its own, and on this site, about a quarter of the tickers hitting that ceiling are empty rows.