Accumulation/Distribution Indicator on the S&P 500: 33 Years of SPY

00The verdict in three lines

• We tested the Accumulation/Distribution line on stocks where it was designed: SPY daily bars from 1993 to October 2026, long only, cash when out, with the three common uses: a bullish divergence, the A/D line crossing its 20-day average, and the Chaikin oscillator crossing zero.
• It did worse than random timing. The two crossing rules earned +0.132% and +0.121% a trade; the same holding periods on random days earned about +0.461%. Their signals ranked below 99% of the random runs.
• Holding SPY turned 1 unit into x29.8; the A/D rules, into x1.74 to x2.11. The divergence fired too rarely to judge. Buried, all three.

01The A/D line on stocks, in one paragraph

The Accumulation/Distribution line adds each day's volume, weighted by where the close sits in the day's range (+1 at the high, −1 at the low), to a running total; the Chaikin oscillator is the 3-day EMA of that line minus its 10-day EMA. A rising line is read as accumulation, a falling one as distribution. The full explanation, and the same three rules on 4-hour Bitcoin candles, are in our Bitcoin A/D test, where none of them proved an edge either.

02The rules, adapted to a stock index

• ADS-DIV: a swing low is a day whose low is below the five days on each side; when a new swing low is lower than the previous one (within 60 trading days) but the A/D line is higher, buy at the next open. Stop at the swing low, target twice the risk, out after 30 trading days. With daily bars only, a day that touches both the stop and the target counts as a stop.
• ADS-MA: buy when the A/D line crosses above its 20-day average, sell to cash when it crosses below.
• ADS-CHO: buy when the Chaikin oscillator crosses above zero, sell when it crosses below.
• Long only, signals at the close, fills at the next open, dividends earned while invested, Treasury-bill interest while in cash. Shorting an index that rose about thirtyfold with dividends would mostly measure that rise, so the yardstick is buy and hold.

We fixed these rules on 22 October 2026, after the Bitcoin result and before computing a single A/D value on SPY. This SPY history was used before by our 200-day moving average test, whose numbers appear below for comparison.

03How we tested it

SPY daily bars, 29 January 1993 to 20 October 2026 (8,489 trading days), from Yahoo Finance, with each day's volume and SPY's dividends; SPY has never split. Costs: 0.04% a round trip, also shown at 0.10%. Periods: 1993 to 2005, 2006 to 2015, and a holdout from 2016 already used by the 200-day test. First-period bar t ≥ 2.0. Control: every trade's holding period, kept at its length, moved together to other days, 500 times. Engine check: the A/D line and the oscillator against a hand calculation, crosses and next-open fills, a dividend inside a trade, swing points, divergences, a gap through the stop, both levels on one day, the 30-day exit, and an independent re-implementation that matched every one of the 1,041 trades (adspy_selftest_result.txt).

04Results: is the A/D line useful on stocks?

Growth of 1 unit from November 1993, each strategy with Treasury-bill interest while in cash, against buy-and-hold SPY with dividends (dashed); log scale. * Holdout already used before this test.

Accumulation/Distribution vs buy and hold

The two crossing rules were in the market about two thirds of the time and kept a small part of its gain. The average-cross version even had a deeper drawdown than holding (end of day, open losses included): it lost 62.3% on closed trades in 2008 (sum of trades), buying rebounds that failed.

05A/D signals or random days?

The control keeps every trade's length and moves the trades to other days. Because SPY rose over the period, almost any long holding period earned something; what matters is whether the A/D chose better ones. It chose worse: for both crossing rules the real result was below nearly every random run.

06By period and by year

The divergence version traded only 76 times in 33 years (32, 24, 20 trades in the three periods), below the 50 per period our protocol asks for; its year-by-year results are in the trade log.

07The gates

Fragile: positive as a sum of trades, negative without the ten best. The holdout gate passed for all three only because SPY rose strongly from 2016; random long periods did better still.

TWhat this shows, and what to take into your own tests

What the results show

• On 33 years of SPY, with real volume and dividends, the A/D line crossing its 20-day average and the Chaikin oscillator crossing zero picked worse moments to be long than chance: the same holding periods placed at random days earned +0.461% and +0.508% a trade, the real signals +0.132% and +0.121%.
• With cash interest when out, 1 unit grew to x1.96 and x1.74 from November 1993, against x29.8 for holding SPY and x16.4 for the 200-day moving average.
• The divergence signal, the use most often taught, fired only 80 times in 33 years of daily bars: 76 trades, too few to judge.

What remains unknown

• Single stocks, where the A/D line is most often drawn. We tested the index ETF; individual shares have their own volume patterns.
• Other averages and oscillator settings. We used the usual 20-day average and Chaikin's 3 and 10.
• Short selling. We tested long only, as most people use the indicator on stocks.

Why it was buried

The two crossing rules bought after days that closed near their highs on volume and sold after days that closed near their lows; one likely reason for the result, which we did not test separately, is that on SPY this often meant buying after a bounce and selling after a dip. They held about 67.9% of the time and still had a worse drawdown than buy and hold, end of day with open losses included (-61.3% for the average-cross version, -55.2% for holding), with -62.3% on closed trades in 2008 alone. The sum of trades was positive for ADS-DIV, ADS-MA and ADS-CHO, and turned negative once the ten best trades of each were removed.

What you can apply to your own tests

• On a market that rises over time, compare a long-only signal with random long periods of the same length. Being long at all is what earns; the question is whether the signal adds to that.
• Compare with simply holding, after costs and with cash interest. Here the indicator kept less than a tenth of the market's gain.
• A volume indicator built for daily stock charts still has to be tested on daily stock charts before you trust it. It did not survive that test here.

A note on the numbers: drawdown figures are measured on closed trades unless a table says it includes open losses. Open losses inside a trade are not counted in a closed-trade drawdown, so real account risk is higher, especially for systems with no stop. Backtests are not live trading.

FAQCommon questions

Is the A/D line useful for stocks?

Not as a timing signal on SPY in our test: both crossing rules ranked below 99% of random long periods of the same length, and the divergence fired only 80 times in 33 years of daily bars.

Does the Chaikin oscillator work on the S&P 500?

Not in our test: buying when it crossed above zero and selling when it crossed below made +0.12% a trade over 455 trades from 1993 to 2026 and grew 1 unit to x1.74 with cash interest, against x29.8 for holding SPY.

Accumulation/Distribution vs buy and hold: which did better?

Buy and hold, by far: x29.8 from November 1993, against x2.11 at best for the A/D rules and x16.4 for the 200-day moving average. The A/D average-cross rule also had a deeper worst drawdown than holding (end of day, open losses included).

Does A/D divergence work on daily stock charts?

It is rare: 80 bullish divergences on SPY in 33 years with our swing rule, 76 trades, too few for our gates; the trades it did take were slightly positive on average but negative without the ten best.

How was the strategy tested?

On SPY daily data with volume and dividends, 1993 to 2026, long only, Treasury-bill interest in cash, 0.04% a round trip, with rules written and dated before the data was run, costs included on every trade, results split into development, validation and holdout periods, a random-entry control as the benchmark, and every variant reported. Read the protocol.

08Check our arithmetic

Code provenance: our own code, written from the specification and Chaikin's published formula.

09Data notes

Source. The SPY daily history (prices, volume, dividends) from the Yahoo Finance page saved for our 200-day test, and the 3-month Treasury-bill rate from FRED. Daily bars cannot show which of a day's high and low came first, so a divergence trade whose stop and target were both inside one day is counted as stopped, the cautious choice. Fills: the next day's open for signals, stops and targets at their level or at a gapped open.

10Corrections

22 Oct 2026, navigation: added a link to the new guide comparing trading costs across all our tests. No result changed. 22 Oct 2026, navigation: added a link to the new cross-market scoreboard, which includes this test. No result changed. 22 Oct 2026, navigation: added a link to the new guide comparing trend following and mean reversion across all our tests. No result changed.

+More tests and guides

• Accumulation distribution indicator
• 200 day moving average strategy
• Golden cross backtest
• Macd crossover strategy backtest
• Trend following vs mean reversion: this test next to the others of its style
• Free prop firm challenge calculator
• What all the backtests found so far
• The backtesting protocol
• How to backtest a trading strategy without fooling yourself
• Backtesting glossary