Adjusted and Unadjusted Prices

A four-for-one stock split turns one share at $400 into four at $100. Nobody lost anything. On an unadjusted chart the series shows a 75 percent overnight collapse, and every indicator computed across that boundary is corrupted.

Adjustment multiplies all prior prices by the split factor so the series is continuous. The same is done for dividends, subtracting the distribution from prior prices so that the ex-dividend step does not read as a decline.

Volume is adjusted too, in the opposite direction, so that a pre-split day's share count is comparable to a post-split one. A provider that adjusts price and not volume produces a chart where a relative volumeVolumeThe number of shares traded in a period - the participation behind a price move.Read the full explanation calculation is wrong by the split factor.

For a long-held dividend payer the divergence compounds. Twenty years of dividend adjustments can place the adjusted price far below the price that was quoted at the time, which is correct as a total-return series and confusing as a price history.

It also means round numbers move. A level at $50 that mattered because it was a round number is drawn at $12.50 on a chart adjusted for a four-for-one split, and it is no longer a round number at all.

The third row is the one platforms rarely offer as an option, which is why long-run technical levels on high-yielding stocks are frequently drawn a long way from where anybody actually transacted.

Adjusting for splits is unambiguous. Adjusting for dividends is not: providers differ on whether to adjust for special dividends, how to treat spin-offs, and whether to assume reinvestment.

Two charts of the same stock over twenty years routinely differ for entirely legitimate reasons. A chart with an unexplained overnight halving is nearly always an unadjusted split rather than an event, and checking the corporate action record resolves it in seconds.

Split adjustment is uncontroversial: a two-for-one split doubled the share count and halved the price on one date, and dividing every earlier price by two removes a discontinuity nobody experienced as a loss. Dividend adjustment is the same idea applied to distributions, and it surprises people because it changes prices that were, at the time, exactly what the tape said.

The logic runs like this. On the ex-dividend date the shares open lower by roughly the payment, because the buyer no longer receives it. Nobody lost that money; it left the share and arrived in the holder's account. A total-return series therefore scales all earlier prices down slightly so that the drop does not appear as a decline, and the result is a series whose percentage changes match what a holder actually earned.

This is also why the yield calculated off an adjusted chart is wrong, and why a long-dated level drawn on an adjusted chart drifts every time a dividend is paid. A level marked at $100 five years ago is not at $100 on the chart today, even though nothing about the level changed.

1. 1Find a known splitPick a security that split in the last few years and look at the split date. A vertical cliff means the series is unadjusted; a continuous line means it is adjusted for at least splits.
2. 2Check a dividend payerOn a security that pays a substantial dividend, compare the chart's price from several years ago against a published historical quote. A gap of a few percent per year of dividends means the series is total-return adjusted.
3. 3Read the provider's noteMost platforms state their convention somewhere in settings or documentation, and most default to split-adjusted and dividend-unadjusted.

The reason to establish this once rather than each time is that the two series answer different questions and neither is a substitute for the other. Percentage-return work, backtests and long-horizon comparisons need the adjusted series. Anything about the actual price at an actual moment, including where a level was and what a filing said the stock traded at, needs the unadjusted one.