OBV & A/D Line

Why not just eyeball the volume bars?

From the previous lesson you know volume matters. But Murphy names a real problem with leaving it at that:

Trying to "eyeball" the vertical volume bars along the bottom of the chart is not always precise enough to detect significant shifts in the volume flow. — John Murphy

The fix: convert the sea of volume bars into a single running cumulative line whose direction you can compare to price. If the price is making new highs but the volume-pressure line is making lower highs, you're seeing in one glance what you'd otherwise have to infer from a dozen vertical bars.

Two classical attempts: OBV (Granville, 1963) and the A/D Line (Chaikin). Same goal, different assumptions about how to assign the day's volume.

OBV — Joseph Granville, 1963

Origin confirmed in Murphy and Kaufman:

Developed and popularized by Joseph Granville in his 1963 book, Granville's New Key to Stock Market Profits. — Murphy

Made famous by Joseph Granville, On-Balance Volume is now a byword in stock analyst circles. — Kaufman

Formula

All-or-nothing, based on the close-vs-prior-close sign:

• If today closes up → add today's entire volume to running total.
• If today closes down → subtract today's entire volume.
• If today closes unchanged → no change.

Kaufman's Table 12.1 walks a six-day manual example: start at 0, add 30 on an up day (30), add 27 on another up day (57), subtract 15 on a down day (42), and so on. The number itself has no economic meaning — it depends on where you started the chart.

Reading it

Murphy is explicit about how to use it:

It is the direction of the OBV line (its trend) that is important and not the actual numbers themselves. The actual OBV values will differ depending on how far back you are charting.

What you're looking for, ordered by signal strength:

1. Divergence between OBV and price — the primary signal
2. OBV leads price — OBV's trendline break precedes price's
3. Direction confirmation — OBV and price aligned = healthy trend

A/D Line — Chaikin's refinement

The all-or-nothing rule of OBV has a problem that's easy to spot. Kaufman nails it:

…a small net change in price can result in all volume being designated to one market direction. Is it reasonable to add all volume to the accumulated index if the S&P gained only a fraction of one point?

The A/D Line (Marc Chaikin's Volume Accumulator) fixes this by weighting each day's volume by where in the range the close landed. Kaufman:

The changes made by Mark Chaikin, which take a percentage of volume based on the relative close of prices within the daily range, seem very sensible and avoid the all-or-nothing technique used in On-Balance Volume.

Formula

Define the Close Location Value (CLV):

CLV runs from −1 (close at the low) to +1 (close at the high). Close at midrange → CLV = 0. Then:

So a huge-volume day that closes exactly in the middle of its range contributes zero — no net buying or selling pressure. That's the key improvement over OBV, which would assign all the day's volume based on the sign of the close-vs-prev-close.

Reading it

Same as OBV — watch direction and divergence. A/D's added resolution catches absorption days that OBV misses. A wide-range day with closing near the midpoint of its own range signals that sellers met buyers and no net pressure emerged, even if the close-vs-prev-close happened to be positive.

OBV vs A/D side-by-side

Toggle between OBV, A/D, and Both, then click "Mark the divergence." The hand-crafted OHLC shows:

• Rally 1: price rises on big volume and closes near the highs → both OBV and A/D climb steeply
• Pullback: volume contracts, OBV drops on the red days, A/D drops proportionally to the weak close locations
• Rally 2: price makes a higher high but on half the volume and with closes landing in the middle of the range — OBV stalls because fewer positive contributions, A/D stalls even more because the weak closes suppress each day's CLV
• Breakdown: when price finally rolls over, both indicators confirm on the downside

The key educational point: A/D catches the weak-close days that OBV registers as "full up." In narrowly-mixed tape, A/D gives you earlier warning.

OBV vs A/D — the conceptual difference

Neither is universally better. OBV is cleaner and easier to reason about; A/D is more informative at price levels where closes cluster near midrange (think earnings days, news-driven reversals). Pros often look at both — divergence on OBV is noteworthy; divergence on both OBV and A/D is a stronger signal.

The close siblings — brief tour

Williams Accumulation/Distribution

Larry Williams + Jim Waters. Per Kaufman (kaufman.txt:15043), it's a different animal — not a cumulative line but a bounded 0–1 oscillator. Their "Daily Raw Figure" (DRF) is 1 when a market opens at the low and closes at the high, and 0 when it opens at the high and closes at the low. Traded like Stochastic with 80/20 bands. If you see "Williams A/D" on a platform, it's this — not Chaikin's A/D Line.

Chaikin Money Flow (CMF)

A bounded 21-day version of A/D. Our reference books don't cover it under that name. If you see "Chaikin Money Flow" in trading software, it's CLV × Volume summed over 21 days divided by 21-day volume sum — bounded roughly between −1 and +1. Same idea as A/D, normalized to a fixed window.

Chaikin Oscillator

MACD applied to the A/D Line: fast EMA − slow EMA of A/D. Zero-line crossovers are the signal. Kaufman uses a 21-day Chaikin Oscillator inside a Bollinger/Keltner squeeze setup (kaufman.txt:12834). The logic is the same as MACD of price — you're just applying it to volume pressure instead.

Money Flow Index (MFI)

Quong & Soudack, 1989. Volume-weighted RSI — same bounded 0–100 scale with 80/20 extremes. Covered under the Wilder-style momentum rubric but factoring in volume. Sometimes called "RSI with volume."

Divergence — the primary use case for OBV/AD

Murphy's summary:

It's when the volume line fails to move in the same direction as prices that a divergence exists and warns of a possible trend reversal.

His specific case studies:

• Intel — "An excellent example of how a bearish divergence between the on balance volume line and the price of Intel correctly warned of a major downturn."
• McDonalds — "The OBV line leading the price higher and warning in advance of the bullish breakout."
• GE, Dec 2010 (Kaufman's example) — "prices continue to move higher throughout December before peaking while the OBV spikes at the end of November before going sideways." OBV topped weeks before price did.

This is the classical "volume leads price" narrative. Whether OBV's lead is universally predictive or cherry-picked from favorable examples is a reasonable question — which brings us to the honest statistical view.

Does it actually work?

Honest answer: the books don't have Bulkowski-grade statistical tests of OBV/AD signals. Bulkowski's Encyclopedia treats OBV as an index entry only, with no hit rates.

Kaufman is skeptical but open:

It is necessary to do a comprehensive test on many markets to find out if the OBV produced better trend results than simply using the price… Tests of interest rates did show improvements using OBV as a substitute for price.

His blunter meta-take on the whole volume-indicator family (kaufman.txt:22218):

By seeing these indicators together, they appear to be a collection of minor manipulations of data… you must be able to interpret everything profitably. Writing clear rules and testing them may be the only way to find out.

Translation: OBV and A/D are useful framings, not proven edges. The value is in the framing — forcing you to watch for volume/price divergence — not in the specific formula. Trade OBV divergence if you want, but don't expect Bulkowski-style win rates from it.

Hidden traps

• Reading absolute OBV levels. They depend on where your chart starts. Only the direction matters.
• The equal-close gap. A day that closes exactly at the prior close contributes nothing to OBV. Rare for individual stocks, common for futures at limit up/down — where OBV misinterprets the light volume as bearish.
• Futures volume is reported a day late per Murphy. OBV on futures is always slightly stale.
• Thin stocks. A/D and OBV both assume real two-way flow. On a stock that trades 50,000 shares/day, a single 100,000-share institutional order distorts the line for weeks.
• All-or-nothing assumption in OBV — the problem A/D was invented to solve. A 0.01% up day contributes the same as a 5% up day. Use A/D when close-location matters.
• Treating OBV divergence as a standalone signal. Murphy uses it as a warning that pairs with price-structure confirmation. Trading every OBV divergence blind doesn't work — some ongoing trends have OBV that lags before catching up.

Quick check

What you now know

• OBV (Granville, 1963): running cumulative sum of signed volume. All-or-nothing — full volume added/subtracted by close-vs-prev-close sign. Absolute value is meaningless; direction and divergence are the signal.
• A/D Line (Chaikin): proportional volume weighted by CLV = (2C − H − L) / (H − L). Same goal, finer resolution — catches midrange-close days that OBV mishandles.
• The core critique of OBV that A/D fixes (Kaufman): "a small net change in price can result in all volume being designated to one market direction."
• Primary use case for both: divergence between indicator and price, warning of trend exhaustion.
• Close cousins: Williams A/D (bounded oscillator, different formula), Chaikin Money Flow (21-day bounded A/D), Chaikin Oscillator (MACD of A/D), MFI (volume-weighted RSI).
• Statistical honesty: no Bulkowski-grade hit rates in our library. Kaufman's take: useful framings, not proven edges.
• The direction-only rule: Murphy is explicit — "it is the direction of the OBV line that is important and not the actual numbers themselves."

Next: VWAP — the volume-weighted average price. The one volume-pressure indicator that professional traders genuinely can't trade without.