The last lesson, and the one that reframes all the others
Every idea in this course so far has been about choosing better. Here is the finding that says choosing better is the smaller lever:
A PSX round trip costs about 3.04 percentage points. At 12 rotations a year, that is 36.5 percentage points of annual drag. Gross median trade -2.89% becomes net -5.93%. Win rate 33.9% becomes 25.4%. [LEDGER: reference_backtest_methodology]
"THE REAL LEVER IS TURNOVER, NOT RANKING."
Sit with the second line. Trading once a month costs you 36.5 points a year before any opinion about any stock is even formed. No ranking improvement in any of our backtests is worth 36.5 points. The largest single lever available to a PSX retail trader is trading less.
The evidence that ranking is the smaller lever
- Absolute Buy Score bands 0-40, 40-55, 55-70, 70-85 all sit at 37-40% on the same 8-week test. The score's level discriminates nothing; only its rank does. [LEDGER:
project_backtest_4y] - A mechanical "rank everything and rotate the top N" product measured, net, over 10.5 years: x1.33 equity, +2.7% CAGR, -77% max drawdown, against KSE buy-and-hold at x3.06 and +11.2%. The recorded conclusion was "do NOT build/sell". [LEDGER:
project_brain_backtest_uat] - Trying to rescue it with stops made it worse: top-5 with an 8% stop measured x0.20, a -14% CAGR with an 88% drawdown. [LEDGER:
project_brain_backtest_uat]
What this course is honestly for
This is why ZEME publishes education rather than a rotation product, and why we say it plainly on the surface where a signal service would be quietest: a mechanical top-N rotation on PSX did not beat holding the index, in our own 10.5-year test, and we are not going to sell you one.
The remaining edge, if you have one, lives in three places this course has already named: selectivity at the very top of the rank, the exit, and not paying the toll twelve times a year.
Evidence · History · Confidence · Counterargument
Evidence, 3.04pp per round trip; 36.5pp/yr at 12 rotations; gross -2.89% to net -5.93%; win 33.9% to 25.4%. [LEDGER: reference_backtest_methodology]
History, 10.5 years, whole panel, applied to real pick histories.
Confidence, high. Cost drag is arithmetic, not a forecast, and it is the most robust finding in the entire ledger.
Counterargument. The fair objection: your personal cost structure may be lower than the modelled 3.04pp, and a genuinely longer holding period (a year, not a month) shrinks the drag toward irrelevance. That is the point rather than a refutation, the lever is turnover, and a lower-turnover investor is pulling it.