ADX trend strength measures how hard a market is trending, not which way it is going. By Welles Wilder's convention, ADX(14) above 25 marks a strong trend and below 20 marks no trend. At the November 4, 2026 close, $QQQ read 30.2 with sellers in control, $IWM sat at 25.1 and falling, while $DIA (17.5) and $SPY (16.4) were in no-trend territory. Across nearly a decade of daily data on all four ETFs, readings above 25 were followed by a wider range of 20-day outcomes. They did not make the trend's direction more likely to continue.
That second finding is the one most ADX tutorials skip. Below is the indicator, the current readings, and the numbers behind that claim.
What ADX, +DI and −DI actually measure
J. Welles Wilder introduced the Average Directional Index and its two companion lines, +DI and −DI, in his 1978 book New Concepts in Technical Trading Systems, according to StockCharts ChartSchool. The three lines answer two separate questions:
- +DI and −DI give direction. When +DI is above −DI, upward movement dominates. When −DI is on top, downward movement dominates.
- ADX gives strength. It rises when one side dominates by a wide margin, whichever side that is. A crash and a melt-up can produce the same ADX.
The 14-period calculation works like this:
- For each day, compute True Range (the largest of high minus low, high minus prior close, and low minus prior close), plus +DM (today's high minus yesterday's high) and −DM (yesterday's low minus today's low). Only the larger of the two directional moves counts, and negatives become zero.
- Smooth all three with Wilder's method. Start with a 14-day sum, then each day subtract one-fourteenth of the running total and add the new value.
- +DI equals 100 times smoothed +DM divided by smoothed TR. −DI uses −DM.
- DX equals 100 times the absolute gap between +DI and −DI, divided by their sum.
- ADX starts as the 14-day average of DX. After that, each ADX equals 13 times the prior ADX plus today's DX, divided by 14.
Per ChartSchool, Wilder treated ADX above 25 as a strong trend and below 20 as no trend, with 20 to 25 a gray zone. Fidelity's Learning Center describes the same thresholds. Because of the heavy smoothing, ChartSchool notes that about 150 periods of data are needed before the values settle.
ADX(14) at the November 4, 2026 close
The readings below come from Nasdaq.com historical data (daily OHLC, with the matching pages for QQQ, IWM and DIA), run through the formula above.
| ETF | Close (Nov 4) | ADX(14) | +DI | −DI | ADX on Oct 7 | Reading |
|---|---|---|---|---|---|---|
| $SPY | $766.81 | 16.4 | 29.4 | 26.2 | 10.5 | No trend, rising |
| $QQQ | $691.95 | 30.2 | 19.7 | 36.8 | 15.8 | Strong, −DI leads |
| $IWM | $284.02 | 25.1 | 27.8 | 23.3 | 38.0 | Strong but fading |
| $DIA | $533.11 | 17.5 | 38.2 | 22.7 | 24.3 | No trend, +DI leads |
$QQQ is the standout. ADX nearly doubled in 20 sessions while −DI took control, and the ETF closed 8.9% below its October 6 closing high of $759.66. $IWM shows the opposite pattern. Its ADX has bled from 38.0 to 25.1 even as +DI regained the lead, so a strong trend is losing force. $DIA has the widest +DI lead of the four, but its ADX of 17.5 means that lead has not lasted long enough to register as a trend.
How the 10-year test was built
The dataset holds 2,513 daily bars per ETF, from November 4, 2016 to November 4, 2026. After a 150-bar warm-up, every session from June 13, 2017 to October 7, 2026 was sorted by its ADX(14) at the close: above 25, 20 to 25, or below 20. Each session's forward 20-day return is the close 20 sessions later divided by that day's close, minus one. That gives 2,343 observations per ETF.
"Direction persisted" means the forward 20-day return had the same sign as the DI lead. That is a gain when +DI led, or a loss when −DI led.
The limits matter. Prices are not dividend-adjusted, so all returns are price-only. There are no costs, taxes or slippage, and readings are taken at the close. Overlapping 20-day windows mean neighboring observations are not independent. Four surviving ETFs and a single window that includes the 2020 crash do not make a universal law.
ADX above 25 vs below 20: the dispersion gap
| ETF / ADX zone | Days | Std dev of 20-day return | 10th pct | 90th pct | Moves beyond ±5% | Direction persisted |
|---|---|---|---|---|---|---|
| $SPY above 25 | 872 | 4.95% | −4.73% | +6.53% | 25.6% | 51.6% |
| $SPY below 20 | 926 | 3.65% | −3.77% | +5.15% | 19.1% | 57.5% |
| $QQQ above 25 | 841 | 6.84% | −6.89% | +9.37% | 43.5% | 48.3% |
| $QQQ below 20 | 988 | 4.80% | −4.77% | +7.03% | 33.6% | 55.4% |
| $IWM above 25 | 619 | 6.82% | −6.98% | +9.73% | 42.0% | 50.6% |
| $IWM below 20 | 1,205 | 6.27% | −6.21% | +7.04% | 35.2% | 48.5% |
| $DIA above 25 | 659 | 5.00% | −4.70% | +6.68% | 29.4% | 46.9% |
| $DIA below 20 | 1,050 | 4.49% | −4.02% | +4.52% | 14.0% | 56.5% |
The dispersion result is consistent. On all four ETFs, the standard deviation of forward 20-day returns was higher when ADX started above 25, and both tails were fatter. On $DIA, the share of 20-day moves bigger than 5% in either direction roughly doubled, from 14.0% to 29.4%. Pooled across the four ETFs, 34.9% of strong-trend readings led to a move beyond ±5%, compared with 25.9% of no-trend readings.
ADX was not a rare signal, either. It sat above 25 on 36.9% of sessions for $SPY and 35.9% for $QQQ. $IWM and $DIA spent the most time below 20, at 51.0% and 44.9% of sessions.
Did the trend direction persist?
This is where the textbook story breaks down. On $SPY, $QQQ and $DIA, the DI direction persisted less often after ADX above 25 than after ADX below 20. On $IWM the two zones came out about even, near a coin flip. Pooled, direction persisted on 49.4% of strong-trend readings and 54.1% of no-trend readings.
The culprit is the downside. Pooling all four ETFs:
- ADX above 25 with +DI leading (1,593 days): the next 20 sessions were higher 62.6% of the time, with an average gain of +0.62%.
- ADX above 25 with −DI leading (1,398 days): the next 20 sessions were lower only 34.3% of the time, and the average 20-day return was +2.15%.
- ADX below 20 with +DI leading (2,484 days): higher 65.6% of the time, with an average gain of +0.98%.
In these index ETFs, a "strong downtrend" reading was more often followed by a rebound than by more selling. On $DIA, 73.2% of those readings were followed by a higher close 20 sessions later. Broad indexes have historically spent most of their time rising and snapped back after selloffs. A trend-strength tool that treats every −DI-led reading the same as a +DI-led one inherits that asymmetry.
Is ADX useful as a filter?
Based on this data, it works as a volatility filter and not as a direction filter.
- Risk sizing. A high ADX flagged wider 20-day ranges on every ETF tested. Traders use that as a cue to widen stops or cut size, which pairs naturally with an ATR trailing stop.
- Not a confirmation stamp. Requiring ADX above 25 before acting on a +DI lead did not improve the hit rate. Pooled, +DI leads were followed by a 20-day gain 62.6% of the time with ADX above 25, compared with 65.6% with ADX below 20.
- Short-side caution. On these index ETFs, −DI-led readings were followed by a 20-day loss only 32.1% to 37.2% of the time in every ADX zone. A high ADX did not fix that, at 34.3%.
For single stocks, sectors or futures, the result could differ. This test covers only four broad US index ETFs. More trend tools are collected in the Analysis section and on the Traders On Trend homepage.
Bull/Bear Verdict
Bull Case: Where it helps: ADX above 25 reliably flagged bigger 20-day swings on $SPY, $QQQ, $IWM and $DIA. That makes it a useful gauge for sizing positions and setting stops, and it separates the two halves of a chart cleanly: DI lines for direction, ADX for strength.
Bear Case: Where it misleads: a high ADX did not make the trend more likely to continue. Pooled persistence was 49.4% above 25 versus 54.1% below 20, and high-ADX readings with −DI leading were followed by gains far more often than losses. Treating "strong trend" as "keep going" did not hold up in this data.
FAQ
What is a good ADX reading for trend strength?
By Wilder's convention, as described by StockCharts and Fidelity, ADX(14) above 25 signals a strong trend, below 20 signals no trend, and 20 to 25 is a gray zone. These are conventions, not proven cut-offs.
Does a rising ADX mean prices are going up?
No. ADX measures strength in either direction. $QQQ's ADX rose from 15.8 on October 7 to 30.2 on November 4, 2026 while the ETF fell and −DI led. Direction comes from comparing +DI and −DI.
Why use 14 periods for ADX?
StockCharts says 14 is the setting Wilder recommended, and it is the default on StockCharts' own charts. All the figures in this article use ADX(14) on daily bars.
Is ADX lagging?
Yes. StockCharts notes that ADX has "a fair amount of lag" because of all the smoothing. In the table above, $IWM's ADX was still at 25.1 on November 4 after falling from 38.0, even though +DI had already retaken the lead.