Weekly Commentary: September 20, 2026
⚠️ Not financial advice. This is auto-generated each week by Anthropic's Claude (an AI model). Brian Beals is not a registered investment advisor, and Anthropic's Claude is not licensed to provide personalized financial advice. The screener is a research and methodology demo, not a recommendation system. Past performance does not predict future results. Do your own research before making any investment decisions.
By Brian Beals. Methodology and code: github.com/brianbeals/sector-rotation-screener. Commentary generated by Anthropic's Claude (claude-opus-4-8).
What the screen said this week
The screen classified the current environment as mid-cycle, driven by industrial production (INDPRO) running at +1.4% year over year, which sits inside the 0.0% to 4.0% band the rule set treats as steady expansion. The yield curve reading of +0.27 (the spread between longer and shorter maturity Treasury yields, a positive value meaning longer yields are higher) reinforced that label. In plain terms, the macro inputs pointed to ongoing but unremarkable growth.
Technology (XLK) topped the composite at 78.4 and was the only sector carrying a Buy signal under this rule set. Its score leaned on a perfect cycle fit of 100.0 and strong relative strength of 74.9, even though its 3-month relative strength (rs3m) was actually negative at -3.10%. Communications (XLC) placed second at 62.9, also with a maxed cycle fit but a much weaker relative strength of 46.5, landing it in Hold territory. No sector triggered the Watch flag this week (which requires cycle fit at or above 80 with relative strength below 45). The bottom of the table was crowded with Avoid signals: Healthcare, Industrials, Consumer Discretionary, Real Estate, Materials, Utilities, and Consumer Staples all scored at or below 40.
Things worth noticing
There is a notable split between recent price momentum and the composite ranking. Energy (XLE) posted the strongest rs3m at +18.21% and Healthcare at +10.95%, yet both landed in Hold or Avoid because the mid-cycle phase does not favor them (cycle fit of 50.0 and 35.0 respectively). Meanwhile XLK, the Buy sector, showed a slightly negative 3-month figure. This is a reminder that the relative strength component and the shorter rs3m readings can point different directions.
XLC's seasonality carries a "thin sample" tag, meaning the historical data behind that 26.7 seasonality score is sparse and worth treating with extra caution. Also worth flagging honestly: since May 2011, this rule set returned +646.90% versus SPY at +654.00%, so it did NOT beat the benchmark net of trading cost. Buy-signal excess forward return was a razor-thin +0.02%, beating SPY only 50% of the time.
Methodology reminder
The composite is a weighted sum: Seasonality 25%, Cycle Fit 40%, and Relative Strength 35%. Lookahead bias in the backtest is controlled using FRED ALFRED data vintages, meaning each historical decision uses only the macro data that was actually available at that time. The backtest figures describe how this specific rule set behaved historically and are not a forecast of future results.