Does it hold up?
Every reading the product shows is measured on three years of weekly data. This is how readings measured that way held up over the year that followed — a year the first measurement never saw. It is published whatever it shows.
Did the direction and the size last?
| Label the reading carried | Readings | Same direction a year later | Consistent within the ranges |
|---|---|---|---|
| Clear | 3,294 | 67% | 62% |
| Tentative | 5,038 | 60% | 74% |
| Not distinguishable from zero | 20,748 | 53% | 83% |
Carry-over at full size: 0.32 (1.00 would mean readings did not shrink at all out of sample).
By force (Clear and Tentative readings)
| Force | Readings | Same direction | Consistent |
|---|---|---|---|
| Interest rates | 2,282 | 72% | 70% |
| Inflation expectations | 1,150 | 61% | 71% |
| U.S. dollar | 1,990 | 65% | 72% |
| Oil and energy | 1,922 | 55% | 64% |
| Credit spreads | 988 | 52% | 73% |
How it was measured
29,080 reading pairs from 494 companies, at 12 start dates every six months (Oct 18, 2019 to Apr 11, 2025). At each start date the five core forces were measured on the 156 weeks before it, exactly as the product measures them, and again on the 52 weeks after it. Pairs from neighbouring start dates share data, so they are not independent; the percentages are descriptive, not a significance test. The study reruns monthly and every run is kept.
This describes the past, and it is not a forecast. Nothing here is a recommendation to buy, sell or hold any security.