Trend
What it measures¶
A moving average is the average of the last so-many closing prices, recalculated every day or week: the 200-day average is the mean of the last 200 daily closes. Averaging removes most of the day-to-day noise and leaves the trend — the slower movement underneath. The Trend tile asks three yes-or-no questions, its votes:
- T1 — is the latest weekly close above the 30-week average, and is that average rising (up by at least 0.5% over four weeks)? This is the test of Stan Weinstein's "Stage 2".
- T2 — is the latest daily close above the 200-day average, and is that average higher than 21 sessions ago? It is the same question as T1 on the daily clock.
- T3 — is the 50-day average above the 200-day average — is the medium-term trend above the long-term one?
The tile is labelled with the Weinstein stage the weekly chart is in: Stage 1 — basing (the price moving sideways around a flat average), Stage 2 — advancing (above a rising average), Stage 3 — top (the average flattening under a choppy price) or Stage 4 — declining (below a falling average). The ADX(14) reading — a measure of how strong a trend is, whatever its direction — adds a caption: below 20 no clear trend, 20 to 25 developing, above 25 strong.
Why it is on the page¶
The long moving average is the oldest tool of trend-following. Its tested value is in reducing deep falls rather than raising returns: Mebane Faber's study of a 10-month average on US shares from 1900 found similar returns with much smaller drawdowns, and an independent 2026 test of Weinstein's stages found that only Stage 2 weeks earned more than the market. The tile places the price in that frame; it does not time anything.
Thresholds, their source and the population they were fitted on¶
| Votes passed | Percent | Verdict |
|---|---|---|
| 3 of 3 | 100% | Strong ✓✓ |
| 2 of 3 | 67% | Good ✓ |
| 1 of 3 | 33% | Weak ✕ |
| 0 of 3 | 0% | Poor ✕✕ |
| T1 and T2 disagree | the pair scores ½ + ½ | Mixed – |
When the weekly and daily readings disagree — the weekly close above a rising 30-week line but the daily close below the 200-day, or the reverse — the tile reads Mixed, names both readings, and scores the pair as half a vote each. The weekly line describes the longer cycle and is not overruled by the daily one, following Alexander Elder's rule that the longer timeframe sets the direction.
- Sources: Stan Weinstein, Secrets for Profiting in Bull and Bear Markets (1988), for the 30-week line and the stages; Faber (2006), "A quantitative approach to tactical asset allocation", for the long average; Mark Minervini for the 50-day over 200-day condition; J. Welles Wilder (1978) for ADX's 20 and 25 lines.
- Fitted on: Faber — the US share market 1900–2005 and later updates; the 2026 stage test — S&P 500 companies' weeks, 1992–2026. The 0.5% "rising" band is the product's own, shared with the Technical page.
- Evidence grade: peer-reviewed and replicated as a drawdown filter; working paper for the stages; practitioner convention for the 50/200 condition and the ADX lines.
A worked example¶
Honeywell International, on daily and weekly bars to 23 October 2026: the weekly close was 2.7% below a 30-week average that fell 1.5% in four weeks (T1 failed), the daily close 4.2% below the 200-day average (T2 failed), and the 50-day average 8.0% below the 200-day (T3 failed): 0 of 3 votes passed · 0% · Poor ✕✕, labelled Stage 4 — declining, ADX 13.1 — no clear trend.
Linde, the industrial-gases company, the same day: the weekly close was 2.0% above a 30-week average that rose only 0.1% in four weeks — flat, so T1 failed — while the daily close was 3.6% above a rising 200-day average (T2 passed). The two clocks disagreed, so the tile read Mixed, with the pair at half a vote each and T3 failed: 1 of 3 votes passed.
What “No verdict” means here¶
Trend needs at least 200 daily bars and 34 weekly bars (30 for the average plus 4 for its slope). With fewer — a company listed recently — the tile reads No verdict — fewer than 200 daily bars (or weekly), shows no rule colour, and its votes leave the strip's count. When the newest daily bar is older than the previous weekday, the tile keeps its verdict and is marked stale with the bar's date.
What the verdict is and is not¶
The verdict describes where the price sits against its own moving averages. It forecasts nothing: a price above a rising average can turn the next week, and the averages lag by design. It is not an instruction to trade the shares, and it says nothing about the company's condition, which the band above judges separately.