Stage Analysis: Weinstein's Four Stages, Tested on 76 Years of the S&P 500

Michael Whitman · Trading since 2017, full-time since 2020 · Published 2026-09-25

Stage analysis is one of the few chart methods built for a person with a job. It asks for one look at a weekly chart and one question: which stage is this in? This page explains the four stages and then measures a mechanical version of the rule on the S&P 500 since 1950.

How it works

The method comes from Stan Weinstein’s book “Secrets for Profiting in Bull and Bear Markets.” ChartMill’s documentation, read on 25 September 2026, summarizes the core of it: Weinstein “uses weekly charts and identifies the current stage based on the direction on the 30 week moving average.”

The 30-week average covers about six months of trading. On a daily chart the nearest match is a 150-day average. A moving average that long moves slowly, which is the point: it filters out the weekly noise and leaves the direction of the larger trend.

Stage 1, basing. After a decline, price stops falling and moves sideways. The 30-week average flattens and price crosses back and forth over it. Sellers are running out, but buyers are not yet in control.

Stage 2, advancing. Price breaks out of the base and holds above a rising 30-week average. This is the only stage the method wants a buyer to own. Summaries of the method usually name a breakout from the base, on heavier trading than usual, as the entry.

Stage 3, topping. The advance stalls. Price chops around a flattening average, and the rises get weaker. Holders are selling into strength.

Stage 4, declining. Price breaks down and stays below a falling 30-week average. The method says not to own it and not to buy it because it looks cheap.

Stages 1 and 3 are the hard part. Both are sideways ranges around a flat average. ChartMill’s page puts it plainly: “the true stage level can only be identified after the next phase starts.”

Relative strength and the market stage

Weinstein also compares each stock with the market. ChartMill names Mansfield relative strength as “a key indicator used in Weinstein’s work”: a line that rises when the stock beats the index and falls when it lags.

The same test can be run on the market itself. A Stage 2 stock in a Stage 4 market has the wind against it, so it makes sense to read the index’s stage first, then the sector’s, then the stock’s.

That makes it a filter more than a signal. It decides which charts deserve attention this week, and leaves the exact entry to a breakout or a pullback rule.

A worked example

Take the S&P 500 in March 2020, using weekly closes. On Friday 20 March 2020 the index closed at 2,304.92. Its 30-week average, the mean of the last 30 weekly closes, was 3,081.47. Four weeks earlier the average had been 3,102.93.

Step 1, the side of the average: 2,304.92 is below 3,081.47, by 25.2%.

Step 2, the direction: 3,081.47 is lower than 3,102.93, so the average is falling.

Step 3, the reading: below a falling average is a Stage 4 reading. The method says to stay out.

Step 4, what happened: 13 weeks later the index was 34.4% higher. The low of that fall was already in, and the rule, built to be slow, said “decline” at the bottom. By the week of 5 June 2020 the index had closed at 3,193.93, back above an average of 3,027.86 that was still edging down, a mixed reading.

The same rule was right in September 2008. On 26 September 2008 the index closed at 1,213.27, below a falling average of 1,315.05, and 13 weeks later it was 28.1% lower. The method gives up the first part of every recovery to stay out of most of every collapse.

The original data

Using the S&P 500’s weekly closes from 25 August 1950 to 18 September 2026, 3,970 weeks, each week was sorted by this site’s mechanical version of the rule: price above or below its 30-week simple average, and the average rising or falling against its value four weeks earlier.

It is a simplification. It cannot tell Stage 1 from Stage 3, so those weeks are grouped as mixed.

Above a rising average, the Stage 2 condition, held in 2,539 weeks, 64.0%. Below a falling average, the Stage 4 condition, held in 865 weeks, 21.8%. The other 566 weeks, 14.3%, were mixed.

The next 13 weeks differed more in risk than in typical return. After an above-and-rising week, the index was higher 13 weeks later 70.3% of the time, with a median change of +2.78%. After a below-and-falling week it was higher 56.5% of the time, with a median of +1.88%. Mixed weeks came in at 66.6% and +3.38%.

The bigger gap is in the bad outcomes. A fall of 10% or more over the next 13 weeks followed 3.2% of the above-and-rising weeks, 80 of 2,526, against 10.1% of the below-and-falling ones, 87 of 865.

For the mixed weeks it was 6.5% of the mixed readings, 37 of 566. The worst 13 weeks after a below-and-falling reading was −38.1%; after an above-and-rising one, −29.3%.

What that suggests: on the index, the 30-week rule has worked mainly as a way to avoid the worst stretches, not as a way to find bigger gains. The reading changed 407 times in 76 years, a little over five times a year.

As of 18 September 2026 the index closed at 7,650.50, above a rising average of 7,298.39. The full weekly file is in the 30-week states CSV.

In this site’s study of 24,971 trading and investing videos, deduplicated by video id, 4 titles name stage analysis, Weinstein or the four stages of the market, from 4 channels. Two reached large audiences, at 1,293,886 and 313,318 views; the other two have 1,139 and 46.

When it fails

The first failure is the lag. A six-month average turns months after price does. The March 2020 example shows the cost: a Stage 4 reading at the low, and a return to the good side only after a large part of the recovery.

The second is whipsaw. Of 120 separate runs above a rising average since 1950, 37 lasted four weeks or fewer. Each short run is a trade entered and closed for little or nothing, before costs.

The third is guessing Stage 1 versus Stage 3 in real time. A flat average with price chopping around it can be a base or a top, and the chart does not say which until it breaks. Buying a “Stage 1” that turns out to be Stage 3 is the method’s most common mistake.

A fourth is applying it to the index and assuming it holds for single stocks. These figures are for the S&P 500. A single stock can gap through its average on earnings, and a thin one can fake a breakout. The volume on the breakout week matters more there.

And a fifth is reading the stages as a forecast. A Stage 2 reading describes the last six months of price. The data above shows it has come with fewer large falls, not that the next week will be up.

Related

The moving average page explains how the 30-week line is built and why a longer average lags more. The relative strength page covers comparing a stock with its market, the second half of Weinstein’s filter. And the market trend page sets out the broader idea of reading direction before choosing trades.

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