The first snapshot of this journal covers March 10 to April 22, 2026 — 43 calendar days. The S&P 500 shed roughly 1.6% over that window and the NASDAQ was essentially flat, barely negative. Not a crash by any measure, but a soft, grinding decline that tends to expose which strategies actually carry edge and which just look good in bull markets.
Across 7 fully-tracked systematic screens (two others, Motley Fool and Graham Net-Net, carry partial data and are excluded from portfolio calculations), the results were split almost perfectly down the middle. Four strategies beat the market. Three didn't.
Best Screen
+6.2%
S&P 500
−1.6%
Worst Screen
−2.1%
The Portfolio Methodology
Before reading the numbers, it is important to understand what they represent. Each strategy is modelled as a $10,000 equal-weight portfolio. At every snapshot, the portfolio rebalances: stocks that dropped off the screen are sold at the last price recorded in the screen data (effectively a 0% return assumption for the exit leg), and new entrants are added at their snapshot price. Retained stocks carry their actual price change between snapshots.
This is a conservative model. In reality you would try to capture the exit price on the day a stock leaves the screen, which could be better or worse than zero. The 0% assumption slightly understates performance when dropped stocks fall further after leaving, and slightly overstates it when they recover. Over a large enough sample it tends to average out.
A strategy isn't a stock pick — it's a repeatable filter. The edge lives in the process, not in any single name.
Portfolio Returns
The table below shows cumulative return through the end of Period 2 (April 27), but given that Period 2 spans only 5 trading days, the performance is almost entirely driven by what happened here in Period 1.
| Strategy | Portfolio Value | Cumulative Return | vs S&P 500 |
|---|---|---|---|
| S8 · CANSLIM Rocket | $10,617 | +6.17% | +7.76% |
| S3 · Lunch | $10,548 | +5.48% | +7.07% |
| S2 · Schloss Dividend | $10,301 | +3.01% | +4.60% |
| S1 · Burry Value Screen | $10,270 | +2.70% | +4.29% |
| S7 · Piotroski F-Score | $9,876 | −1.24% | +0.35% |
| S9 · Quality Growth | $9,886 | −1.14% | +0.45% |
| S6 · Magic Formula | $9,791 | −2.09% | −0.50% |
| Benchmark · S&P 500 | $9,841 | −1.59% | — |
| Benchmark · NASDAQ-100 | $9,942 | −0.59% | — |
CANSLIM Rocket and the Lunch screen were the standout performers. Both target momentum and earnings growth — precisely the characteristics that tend to hold up even when the broader index softens, because growth stocks with accelerating earnings attract buying regardless of the macro backdrop.
The Magic Formula was the only strategy to meaningfully underperform the S&P 500. At −2.09% against −1.59% for the index, the spread is small, but combined with a retention rate of just 33% (the screen was essentially rebuilt from scratch between the two dates), it raises questions about whether the screen is genuinely identifying durable value or chasing cheap multiples in deteriorating businesses.
Retention: How Stable Were the Screens?
Retention rate tells you how much of the portfolio survives unchanged from one snapshot to the next. High retention means the strategy is selecting genuinely durable businesses. Low retention means the screen is sensitive to short-term fluctuations in the underlying metrics — or that the market has rotated hard enough to flush the screen entirely.
The tension here is instructive. CANSLIM Rocket had a retention rate of only 57.9% — nearly half the portfolio turned over — yet it was the best performer. This is the nature of momentum strategies: they are designed to rotate quickly into whatever is working and out of whatever isn't. High turnover is a feature, not a bug, provided the replacement stocks are better than what was sold.
By contrast, Quality Growth (S9) had 100% retention and still underperformed the market slightly at −1.14%. The screen is holding genuinely high-quality businesses, but quality alone doesn't protect against a broad market decline. These companies simply declined with everything else.
The Piotroski F-Score at 25% retention is the most concerning data point. Three of four original stocks were replaced by an entirely new set of 15. That is not rebalancing — that is almost starting from scratch. The F-Score is sensitive to quarterly fundamental changes, so large turnover is expected, but this level of churn over 43 days suggests the universe of passing stocks is narrow and volatile.
What This Period Tells Us
Four observations stand out from Period 1:
1. Momentum beat value in a soft decline. CANSLIM and Lunch (both oriented toward earnings acceleration and relative strength) outperformed value-focused screens like Magic Formula and Piotroski. This is historically consistent — value often underperforms in the early stages of a correction, recovering later as valuations compress further.
2. High retention correlates with predictability, not necessarily returns. The stable screens (S1, S9) delivered consistent, moderate results. The volatile screens (S8, S7) swung harder in both directions. Know which type you're running.
3. The sample is too small to draw conclusions. Seven strategies over 43 days is not a backtest — it is a single data point. The purpose of this journal is to accumulate enough observations across different market regimes to say something meaningful. Period 1 is chapter one of a longer story.
4. Exited stocks assumed 0%. This is a known conservatism in the model. Stocks that leave a momentum screen often continue declining; stocks that leave a quality screen often stabilise. The direction of bias differs by strategy type, which adds further complexity to cross-strategy comparisons.