Moving Average Indicator: SMA, EMA, WMA and Hull Compared, With Their Lag Measured on Real Prices
What a moving average shows and what it cannot, the formulas of the main types, how many bars each one lags on daily Bitcoin and the S&P 500 ETF, and what the 200-day average actually did on both.
On this page
A moving average is the average price of the last n bars, recalculated on every new bar, so it slides along the chart as a smoothed version of price. It does not predict anything: it tells you the direction of the trend over the period you chose, and it always lags price because it is built from past bars. The main choices are the length and the weighting. A 20-day simple average follows short swings, a 200-day average the long trend; an exponential or weighted average gives more weight to recent bars and turns sooner. Below we measure that lag on real daily prices, so you can see what each choice costs.
SMA, EMA and WMA of the same length on daily Bitcoin
- Price, daily candles (hollow up, filled down)
- SMA(20) (dashed line)
- EMA(20) (solid line, second color)
- WMA(20) (solid line)
The main types of moving average
| Type | How it weights the bars | Effect |
|---|---|---|
| Simple (SMA) | All n closes equally | Smooth, the most lag for a given length |
| Exponential (EMA) | Each bar by a factor 2 / (n + 1), older bars decaying geometrically | Turns sooner than an SMA of the same length |
| Weighted (WMA) | Linearly, from 1 for the oldest bar to n for the latest | Lagged as little as the EMA or less in our measurement |
| Smoothed (SMMA) | Like an EMA with a factor 1 / n | Very slow; used by Wilder's indicators and the Alligator |
| Hull (HMA) | Weighted averages combined to cancel most of the lag | Fastest, but noisier; see the Hull moving average |
Other designs answer the same trade-off between smoothness and lag in other ways: the Arnaud Legoux moving average, the McGinley Dynamic and the volume-weighted average.
The difference between the types is only in how much each past bar counts. Drawn for the same length, the weights look like this:
How SMA, WMA and EMA weight the last bars
The EMA never quite drops a bar: each one keeps a weight that shrinks by the same factor every bar, which is why platforms that start it differently agree only after a few lengths of data. The SMA drops the oldest bar all at once, so a large move leaving the window can turn the average even on a quiet day.
How much each average lags, measured
We measured the lag of each type on our two daily series, Bitcoin and the S&P 500 ETF (dividend-adjusted prices), to October 28, 2026, starting after three lengths of warm-up: the 20-bar lines from May 7, 2025 on Bitcoin and September 4, 2024 on the ETF, the 50-bar lines from August 5, 2025 and January 14, 2025. Lag here is the shift k, from 0 to 40 bars, that minimizes the mean of |averaget / closet−k − 1| over that period, the average gap in percent between the average today and the close k bars ago. The result can move by one bar with the way the gap is measured: with the gap in price instead of percent, the Bitcoin SMA(50) comes out at 25 bars, the WMA(20) at 5 and the WMA(50) at 13. We also counted how often the close crossed each average over the same period, a rough count of the signals it would give: a crossing is a close on the other side of the average from the close the bar before.
Lag of each average, in bars
- Bitcoin
- S&P 500 ETF
Close crossings of each average
- Bitcoin
- S&P 500 ETF
A simple average lags by about half its length, which matches its construction: it is centered on the middle of its window. The weighted and exponential forms cut that by about a fifth to a half, and the Hull by about 70% or more. The cost shows in the crossing counts: on 20 bars, the less an average lagged, the more often price crossed it; on 50 bars the same held for the WMA and the Hull, but the EMA was crossed slightly less often than the SMA (30 against 31 on Bitcoin, 29 against 31 on the ETF). On the same 20-bar length the Hull was crossed about one and a half to two times as often as the simple average (118 against 61 on Bitcoin, 109 against 70 on the ETF). There is no free reduction of lag; each step toward speed buys more signals in ranges.
Common periods: 20, 50 and 200
StockCharts describes shorter averages of 5 to 20 periods as suited to short-term trends, 20 to 60 periods to medium-term trends, and 100 periods or more to long-term investing, and calls the 200-day average perhaps the most popular, with the 50-day popular for the medium-term trend. There is nothing special in the numbers themselves; they are popular because many people watch them.
200-day moving average on the S&P 500 ETF
- Price, daily candles (hollow up, filled down)
- SMA(200) of the close (solid line)
The chart shows the 200-day simple average on the S&P 500 ETF from October 30, 2025 to October 28, 2026. Over the 600 daily bars of this chart series, from the first 200-day value on March 26, 2025, when the ETF was below its average, the close crossed it three times, the first of them before the window shown on the chart:
- up on May 12, 2025 at 573.48; the ETF stayed above for ten months, closing at 645.30 when it crossed down on March 20, 2026 (+12.52%), after a high close of 690.10;
- down on March 20, 2026 at 645.30; the lowest close after it was 628.78, and the ETF crossed back up on April 8 at 672.60, 4.23% higher than where it had crossed down;
- up on April 8, 2026 at 672.60; still above on October 28, 2026 at 772.31 (+14.82%).
On daily Bitcoin, over the 600 bars of its series, from the first 200-day value on September 23, 2025, the same average gave eight crossings. Six came in two weeks, from October 17 to October 31, 2025, as the close hovered around the average, between about 106,400 and 110,100 at the crossings. The seventh, down on November 3, 2025 at 106,583.04, was followed by a fall to a lowest close of 58,624.71 before the eighth, up on August 19, 2026 at 69,334.79, 34.95% below the crossing down. On October 28, 2026 Bitcoin closed at 81,343.50, 17.32% above that crossing up. One average, two markets: on the ETF, two long trends and one whipsaw; on Bitcoin, six whipsaws in a fortnight before it caught the decline. The second chart shows Bitcoin from October 14, 2025 to June 20, 2026: the six crossings of late October, the crossing down of November 3 and most of the fall that followed.
200-day moving average on daily Bitcoin
- Price, daily candles (hollow up, filled down)
- SMA(200) of the close (solid line)
Moving averages as support and resistance
StockCharts notes that moving averages can act as support in an uptrend and resistance in a downtrend, and advises treating them as zones rather than exact levels. Price often overshoots a widely watched average before turning, or slices through it; a level that held twice is not a guarantee of a third time. Our bench does not measure moving average bounces.
Crossovers and ribbons
Two averages of different lengths give a crossover signal when the faster crosses the slower. StockCharts calls the shorter average crossing above the longer one a golden cross and the cross below a death cross, the best known pair being the 50-day and 200-day averages; see moving average crossovers. A moving average ribbon plots a set of averages of increasing length on the same chart, for example exponential averages from 20 to 55 bars. When the ribbon fans out in order, fastest on the outside, the trend is strong; when it twists and the averages cross each other, the trend is pausing or turning. The Guppy multiple moving average is a ribbon split into a short and a long group, and its three states are the clearest way to see the idea:
The three states of the GMMA groups
Does it work?
We have no long-run measurement of moving average rules: our bench tests chart patterns detected by rule against random entries, not indicators (see our methodology). The lag and crossing counts above describe two markets over about two years; they show the trade-off between types and lengths, not which one makes money.
If you want to study moving average systems with backtests, Steve Burns' courses on New Trader University cover them: Moving Averages 101, an introduction in an 11-part video course, and Moving Average Signals, which presents backtests of 10 moving average signals across ten stocks. They are affiliate links, and we have not taken the courses.
For the full family, see the list of trend indicators; for averages inside other tools, Bollinger Bands and the MACD.
Frequently asked questions
How do you read a moving average?
By its slope and by where price sits relative to it. A rising average with price above it describes an uptrend over that period; a falling one with price below, a downtrend. A flat average that price keeps crossing describes a range.
What is the best moving average indicator?
None is best in general: each type trades lag against noise. On our data, the Hull average lagged least and was crossed most often; the simple average lagged most, and at 20 bars it was also crossed least. Pick the one whose behavior fits how long you hold trades.
What are the most common moving average periods?
20 for the short term, 50 for the medium term and 200 for the long term are the ones most charts show; StockCharts calls the 200-day perhaps the most popular. For daily charts, 5 to 20 bars covers short-term trends and 100 or more long-term ones.
What does a 20-period moving average tell you?
The average close of the last 20 bars, about one trading month on a daily chart of stocks. On our two daily series a 20-bar simple average lagged price by about 10 bars.
Is the EMA better than the SMA?
It reacts sooner, which on our data cut the lag of a 20-bar average from about 10 bars to 6 or 7, at the cost of a few more crossings. Whether that is better depends on whether early signals or fewer false ones matter more to you.
What is a moving average ribbon?
Several moving averages of increasing length on the same chart. Averages fanning out in order describe a strong trend; averages twisting through each other describe a pause or a turn. The GMMA is a well known ribbon.
