Moving averages and trend: filters, the 50/200 and golden crosses

A moving average smooths price into a single line, and the direction of that line is the simplest definition of a trend there is. If the 200-day average is rising and price is above it, most chart readers will call the stock an uptrend without drawing anything else. That simplicity is why moving averages are on nearly every chart, in nearly every screener and in a good share of the world's systematic trading rules.

This guide is about using moving averages to read and filter trends: what the slope and the price's position tell you, why the 50-day and 200-day lines became the standard pair, and what the famous golden and death crosses actually did on real data from 2023 to 2026. The arithmetic of the averages themselves is covered in the Simple moving average (SMA): formula, crossovers and limits and Exponential moving average (EMA): formula and how to use it guides.

How a moving average describes a trend

A moving average answers one question: what has the typical price been over the last n periods? Three readings follow from it.

1. The slope. A rising average means recent prices are higher than the prices dropping out of the window. A falling one means the opposite. The slope of a long average is the cleanest single description of the long-term trend.

2. Price relative to the average. A close above the average means the latest price is higher than the recent typical price, a close below means it is lower. Many traders combine the two: price above a rising average is an uptrend, price below a falling average is a downtrend, and the mixed cases are transitions.

3. The order of several averages. When a short average sits above a medium one, which sits above a long one, and all three rise, the trend is aligned on every timescale. When the order is reversed, the downtrend is aligned. Tangled averages that cross back and forth mean there is no clear trend.

Why averages lag, and by how much

A simple moving average of n periods is centred on the middle of its window, so it reflects prices from about (n − 1) ÷ 2 periods ago. A 50-day average lags price by about 25 sessions and a 200-day by about 100, close to five months. Exponential averages weight recent prices more and lag less, at the cost of reacting to more noise. No average can turn before price does, so every signal it gives is a confirmation of something that already started.

Why the 50-day and 200-day became the standard

There is nothing special in the numbers 50 and 200. They became standard because they are round, because 200 trading days is close to a calendar year of sessions and 50 close to a quarter, and because financial media and charting platforms adopted them decades ago. Once enough people watch the same lines, the lines matter for the reason any watched level does: orders cluster around them (see Support and resistance: why price levels form and how to use them).

On weekly charts the equivalents are the 10-week and 40-week averages. Ten weeks is 50 sessions and 40 weeks is 200, so a weekly chart with those two lines shows almost the same picture as a daily chart with the 50 and the 200, with much less day-to-day noise. On monthly charts, the 10-month average plays the 200-day role; Mebane Faber's widely cited 2007 paper used it as a rule for moving between stocks and cash.

Golden crosses and death crosses

A golden cross is the 50-day average crossing above the 200-day average. A death cross is the 50-day crossing below it. Both are headlines whenever a major index makes one, and both are slow by construction: the 50-day has to climb (or fall) through a 200-day that barely moves, which takes weeks after price has already turned.

SPY on a weekly chart, 2021 to 2026

The weekly chart shows the three crosses since the 40-week line starts:

• Golden cross, week of January 23, 2023. SPY closed that week at $405.68, about 16.5% above its low of the prior 60 weeks. Twenty-six weeks later it was 12.6% higher.
• Death cross, week of April 7, 2025. This came in the week of the April 2025 sell-off low. SPY closed the week at $533.94. Thirteen weeks later it was 16.8% higher, and 26 weeks later 22.3% higher. Anyone treating the cross as a sell signal sold close to the bottom.
• Golden cross, week of June 23, 2025. SPY closed the week at $614.91, already about 28% above the April low. Twenty-six weeks later it was 12.3% higher.

Two of the three were followed by gains in line with a rising market, and one was a costly false alarm. That pattern is typical of a strong bull market: the crosses arrive late, and the death crosses come after much of the damage is done.

Fifteen stocks and ETFs on daily charts

To see more than one chart, we looked at daily 50/200 crosses for 15 large US stocks and ETFs (SPY, QQQ, DIA, IWM, IYT, AAPL, MSFT, NVDA, TSLA, AMZN, META, GOOGL, AMD, JPM and XOM) from August 19, 2025, the first day our two-year daily data supports a 200-day average, to October 29, 2026. There were 14 crosses, seven of each kind. Ten had 60 sessions of follow-up.

What the table shows:

1. The crosses were late. Golden crosses came on average 31% above the prior 60-session low (all seven, including the two too recent to score). Death crosses came on average about 18% below the prior 60-session high. Most of each move had already happened.
2. Death crosses did not predict declines here. Four of the five with follow-up were followed by gains, three of them by more than 12%. Only MSFT's January 2026 cross was followed by a further fall.
3. Golden crosses were followed by gains, and so was almost everything. All five rose over the next 60 sessions, by 8.8% on average. Over the same period SPY rose in 88% of all 60-session windows, by 3.8% on average. In a rising market, "bullish" signals look good whether or not they add anything.

Whipsaws: when the trend filter flips back and forth

The weakness of every moving-average rule is the whipsaw: price crosses the average, reverses, crosses back, and the rule pays a small loss each time. It happens most around flat averages in sideways markets.

Over 302 sessions from August 2025 to October 2026, Tesla's close crossed its 200-day average 18 times, and spent only 56% of sessions above it. Over the same sessions SPY crossed its 200-day twice and closed above it 96% of the time. A rule of "own it above the 200-day, sell below" would have been almost costless on SPY and a series of small losses on TSLA, whose volatility kept dragging it across a line that was going nowhere for months.

Common ways to reduce whipsaws, each with a cost:

• A band. Act only on a close some distance beyond the average, say 2% or a multiple of the Average True Range (ATR): volatility, stops and position size. Fewer false signals, later real ones.
• Time confirmation. Require two or three closes beyond the average, or use weekly closes instead of daily ones.
• Slope confirmation. Act on price crossing the average only when the average itself is sloping in the same direction.
• Longer averages. Fewer crosses, more lag.

The Average Directional Index (ADX): measuring trend strength is another option: it measures trend strength directly, and a low reading warns that moving-average signals are likely to whipsaw.

What the research says

Academic tests of moving-average rules go back decades. Brock, Lakonishok and LeBaron (1992) tested simple rules, including price against a 50-day and 200-day average, on the Dow Jones Industrial Average from 1897 to 1986 and found that returns after buy signals were higher than after sell signals. Sullivan, Timmermann and White (1999) then showed how much of that kind of result can come from data snooping, testing many rules on one history, and found that the best rules' advantage did not hold up in the following decade. Faber's 2007 paper argued that a 10-month average's main benefit is lower volatility and smaller drawdowns, not higher returns.

The fair summary is that moving-average trend filters have mostly helped by keeping investors out of the worst parts of long bear markets, and have mostly cost money through whipsaws and late re-entry the rest of the time. Whether the trade is worth it depends on the market, the costs and the investor.

Using moving averages to read trends, step by step

1. Pick the horizon. 20-day for short swings, 50-day for the intermediate trend, 200-day (or 40-week) for the long-term trend. Match the average to the trend you actually trade.
2. Read the slope first. Rising, falling or flat. A flat average says "no trend," and crosses of it carry little information.
3. Then the position. Price above or below, and how far. A stock 30% above its 200-day is in an uptrend and also stretched; mean reversion becomes more likely, not certain.
4. Then the alignment. Are the short, medium and long averages stacked in order?
5. Use the average as a reference, not a wall. Pullbacks in strong trends often stop near the 20-day or 50-day average, which is why traders watch them as dynamic support. They also slice through them regularly.
6. Write the rule down and test it. On a range of symbols and years, including bear markets, with costs. Finabase's backtester, which runs an indicator strategy over a year of real prices, is a place to start; a year is not a full market cycle.

Moving-average trend tools compared

How Finabase shows moving averages and trend

On a Finabase chart, the default layout already carries two simple moving averages chosen for the range you are viewing, 50 and 200 on one-year, two-year and year-to-date charts. You can add your own SMA or EMA studies with any period; the two-year SPY chart with the 50 and 200 is a good place to start. Finabase computes the averages with TA-Lib, the same library used for every indicator on the chart.

The technical gauge on each symbol page checks the latest close against a set of simple and exponential averages from 5 to 200 periods and counts a close above each one as bullish and below it as bearish, a quick read of how aligned the trend is across timescales. The Scanner lists the stocks whose close has just crossed one of those averages, with the date the signal fired, and the backtester lets you test a moving-average rule before trusting it.

Frequently asked questions

What does it mean when a stock is above its 200-day moving average?

It means today's price is higher than the average close of roughly the past year. Combined with a rising 200-day line, it is the most common definition of a long-term uptrend. It says nothing about whether the trend will continue.

Is a golden cross a buy signal?

It is a confirmation that the intermediate trend has turned up relative to the long-term trend, and it usually arrives well after the low. In our sample golden crosses came, on average, 31% above the prior 60-session low. Whether acting on it helps depends on the market; test it on many symbols and years before relying on it.

Is a death cross a reliable sell signal?

Not on its own. In our 2025–2026 sample, four of five death crosses with 60 sessions of follow-up were followed by higher prices, and SPY's weekly death cross arrived the week of the April 2025 low. Death crosses do occur early in some long bear markets, which is why trend followers use them, accepting many false alarms.

Should I use a simple or an exponential moving average?

An exponential average reacts faster because it weights recent prices more, so it gives earlier signals and more false ones. A simple average is steadier and is what most published 50-day and 200-day figures use. Neither is better in general; the Exponential moving average (EMA): formula and how to use it guide compares them.

What is the best moving average for identifying trends?

There is no best one. The 200-day (or 40-week) is the standard long-term filter, the 50-day the standard intermediate one, and the 20-day a common short-term one. Choose the period that matches the length of the trends you care about, and expect a longer period to be steadier and later.

Summary

Moving averages turn a noisy price series into a trend you can read at a glance: the slope says which way, the position of price says how the latest move compares, and the order of several averages says whether timescales agree. The 50-day and 200-day became the standard because everyone watches them. Their crossovers are slow confirmations rather than forecasts, as SPY's weekly death cross in the week of the April 2025 low and our 15-stock sample both showed. Use averages as filters and references, expect whipsaws around flat lines, and test any rule across full market cycles before trusting it.