EMA vs Hull Moving Average

Michael Whitman · Trading since 2017, full-time since 2020 · Published 2026-09-08 · Updated 2026-10-19 · Reviewed 2026-10-04

One is the standard responsive average. The other was built to remove lag by projecting the average forward. On SPY the projection does what it promises, and it also does something an ordinary average does not: it runs ahead of every price in its own window.

What each one is

The exponential moving average (EMA) gives each new close a weight of 2/(N+1) and carries the rest from the day before, starting from a simple average. The exponential moving average covers it.

The hull moving average, from Alan Hull, is a weighted average of a projection. For length 20: take twice the 10-day weighted average, subtract the 20-day weighted average, then take a 4-day weighted average of that, 4 being the whole-number square root of 20. Hull moving average covers it.

The subtraction is the point. Twice the short average minus the long one extends the recent slope, so the hull leans toward where price has been heading rather than where it has been.

Where they differ

Where the line can sit. An exponential average is a blend of past closes, so it stays near them. The hull’s projection can land above the highest recent close in a rally or below the lowest in a fall.

How often each turns. An exponential line turns exactly when a close crosses it, because each new value moves toward that close. The hull’s turns come from its own smoothed slope, not from price crossing it.

How close each stays to price. On SPY the hull was a median 0.78% from the close and the 20-day EMA 1.32%.

What each is good for. The exponential line works as a level price can be above or below. The hull’s level can be a price nobody traded, so its direction is the more useful reading.

Which one to use: EMA or Hull Moving Average?

Use the exponential average when you need a line to measure price against, such as above or below a trend line, because it stays among the closes it averages.

Use the hull when you want its turns and can ignore its level. On SPY its turns were less frequent and less often reversed within a few sessions than the EMA’s.

Use the exponential average for stops. A stop placed on a hull line can sit beyond every recent price, which no ordinary average does.

And test either at the length you will trade. Both behave differently at 9, 20 and 50 sessions.

Turns and overshoot on SPY

The test. SPY daily closes from 3 January 1994 to 2 October 2026, both lines at length 20. A turn is a session when the line’s direction changes from the last session it moved, ignoring flat days.

The hull turned less often. 915 turns (27.94 a year) against 1,057 for the EMA (32.28 a year). The median gap between turns was 8 sessions for the hull and 3 for the EMA. Of the gaps between consecutive turns, 20.13% of the hull’s and 51.33% of the EMA’s were 3 sessions or shorter.

The hull overshot. It sat above every one of the last 20 closes on 720 sessions (8.73%), 549 of them while rising, and below every one on 57 (0.69%). The exponential average was never above or below all 20 closes on any session measured.

What an overshoot looks like on the chart

A recent case. On 28 April 2026 SPY’s highest close of the last 20 sessions was $715.17. The hull stood at $720.39, above every one of them, while the close was $711.69 and the 20-day EMA $694.47. The hull was showing a price SPY had not closed at; the EMA was well inside the range.

The biggest overshoot was on 28 October 2002, when the hull stood 1.39% above the highest close of its window. Reading the hull as a level on those days meant reading a projection.

This is why the hull’s lower lag is real but not free. It reaches turns sooner by extending the slope, and the same extension carries it past price when the slope stalls.

The original data

The sample. Closing prices of SPY from Yahoo Finance, no dividends, loaded from the fund’s first session in January 1993 and measured over 1994 to October 2026. EMA: multiplier 2 / 21, seeded with the first 20-close simple average. Hull: as defined above, from Alan Hull’s published formula, weighted averages with weights 1 to N.

The headline figure. The hull sat above all 20 closes of its window on 8.73% of sessions; the exponential line on 0%.

The file. Every session’s close, both lines, the highest and lowest close of the last 20, and each line’s turns are in the SPY hull and EMA file.

Limits. No test here asks whether a turn led to further movement the same way, and no trading rule is scored. A turn counted here is a change of direction, not a signal tested for anything.

When it fails

The failure is a stop placed on the hull line. In a fast rally the hull rises above every recent close, the trader trails a stop just under it, and the first ordinary down day closes below a line that was sitting above the market. The stop fires on a price the market never traded, and it fires on the noise the hull was supposed to filter. On SPY the hull was in that position on 720 sessions. A stop belongs at a price that traded, or on a line that stays among them.

The second failure is expecting a faster line to turn more often. On SPY the hull turned less often than the 20-day EMA.

A third is reading the hull’s level as support. It can be a projection.

A fourth is comparing a 9-day hull with a 20-day EMA. Match the lengths first.

And a fifth is swapping one line for the other after a few bad weeks. Neither record then means anything.

Related

The exponential moving average covers the standard responsive line. Hull moving average covers the projected one. And moving average crossover covers the signal both are usually traded through.

This page is educational, not financial advice. Test every idea on your own charts before risking money.