Home / Data Quality / Adjusted Prices
DataIntermediate

Adjusted Prices

Adjusted prices are historical prices rescaled to remove the mechanical jumps caused by splits, bonuses and dividends, producing a continuous series suitable for computing returns and indicators — but the standard back-adjustment method rescales the past using factors only known later, so it can quietly embed look-ahead into a backtest.

Quick answer: Adjusted prices are historical prices rescaled to remove the mechanical jumps caused by splits, bonuses and dividends, producing a continuous series suitable for computing returns and indicators — but the standard back-adjustment method rescales the past using factors only known later, so it can quietly embed look-ahead into a backtest.

In simple words

Adjusted prices are what you get after smoothing out the artificial jumps that corporate actions leave in raw prices. They make a stock's history continuous, so returns and indicators behave sensibly across a split or a big dividend. The subtlety is that the usual way of adjusting rewrites the entire past using information about actions that had not happened yet, which can sneak the future into your backtest.

Purpose

Adjusted prices exist so that a return series is economically meaningful across corporate actions; the reason to study them carefully is that the convenient adjusted feed most people use carries a built-in look-ahead trap.

Visual explanation

Adjusted Prices

Raw prices plus a dated corporate-action master flowing through split, bonus and dividend adjustment into a continuous, point-in-time series.

Data PipelineSource / FeedIngestNormaliseCorp-actionAdjustValidateStoreServe

Professional explanation

What adjustment does

Raw exchange prices contain mechanical discontinuities: a split divides the price, a dividend drops it on the ex-date. Adjustment removes these so that the percentage change from one bar to the next reflects an investor's real experience rather than a bookkeeping event. A split or bonus adjustment simply rescales all prior prices by the ratio; a dividend adjustment rescales prior prices by a factor that credits the payout back, turning a price series into a total-return series. The output is a continuous curve on which a moving average, a volatility estimate or a momentum signal is not corrupted by the jump.

Back-adjustment and the reference point

The most common technique is back-adjustment: fix the most recent price as the reference and rescale the entire earlier history so it lines up. This keeps today's price equal to the real quote, which is convenient, but it means every historical price is expressed in today's adjusted terms. An alternative is forward-adjustment, which fixes an early reference and rescales later prices, so the oldest prices are real and recent ones are scaled. Which end is real matters for anything that compares an adjusted price to an absolute level, because only one end of the series equals what actually traded.

Why adjusted series can leak the future

The look-ahead danger is structural. A fully back-adjusted series applied today uses the cumulative product of every split and dividend factor across the whole history, including actions that occurred after any given historical date. So a signal computed on an early bar of a back-adjusted series is implicitly scaled by knowledge of splits and dividends that had not yet been announced at that time. For scale-invariant logic — percentage returns, ratio-based indicators — this usually does not change the decision. For price-level logic — a fixed rupee stop, a round-number breakout, a channel in absolute terms — it can materially alter what the backtest does.

Point-in-time adjustment as the fix

The disciplined remedy is point-in-time adjustment: for each date in the backtest, apply only the adjustment factors that were known as of that date, so early bars are unaffected by later actions. This reproduces what a live trader would actually have seen on their screen at each moment, and it is the correct basis for any strategy whose logic depends on price levels. It is more work, because you cannot precompute a single adjusted series; you effectively re-derive the series as of each decision date, or store raw prices plus a dated action master and adjust on the fly.

Adjusted versus unadjusted for different tasks

Neither series is universally correct; the right choice depends on the task. Return and performance calculations, indicator computation and cross-sectional comparison generally want an adjusted, ideally total-return series so that a dividend is not mistaken for a loss. Execution modelling, position sizing in shares, and anything involving the actual rupee price you would have paid want the raw price that really traded, because that is what your broker filled and what STT and brokerage applied to. Mixing them — sizing on raw prices but signalling on back-adjusted ones — is a common and hard-to-spot inconsistency.

Formula

AdjustedClose = RawClose × CumulativeFactor, where split factor = 1 ÷ r and dividend factor = 1 − (D ÷ PrevClose)

RawClose = the actual traded close; CumulativeFactor = product of all adjustment factors applied for actions on or after that date under back-adjustment; r = split or bonus ratio (e.g. 2 for a 1:1 bonus or 2-for-1 split); D = dividend per share; PrevClose = the close on the day before the ex-date. Factors compound multiplicatively across multiple actions.

Back-adjusted vs Point-in-time adjusted

AspectBack-adjustedPoint-in-time
Reference pointMost recent price is realEach date uses only factors known then
Look-ahead riskPresent for price-level rulesNone
ConvenienceOne precomputed seriesRe-derived per decision date
Good forReturns, ratio indicatorsFixed-price stops, absolute levels
Old prices equal reality?No — scaled to todayYes — as seen at the time

Practical example

Illustrative example (Indian market)

An NSE stock does a 1:1 bonus, so its raw price falls from Rs 2,000 to Rs 1,000 on the ex-date. In a back-adjusted series you halve everything before the ex-date, so the historical Rs 2,000 becomes Rs 1,000 and the curve is smooth. Now suppose your strategy uses a fixed breakout level of Rs 1,050. On the back-adjusted series, prices that were really Rs 2,100 before the bonus now read Rs 1,050, so your level is triggered on bars where the true price was double your intended threshold — the backtest is acting on knowledge of the future bonus. A point-in-time series leaves the pre-bonus price at Rs 2,100, your Rs 1,050 level is never wrongly hit, and the backtest matches what a trader watching the real screen would have done.

Most free Indian price feeds distribute back-adjusted series without saying so, and they often adjust for splits and bonuses but not consistently for dividends. A backtester who assumes the feed is raw, or who runs fixed-rupee stops on a silently back-adjusted series, will get results that cannot be reproduced live. Storing NSE raw closes alongside a dated corporate-action master is the only way to control the adjustment yourself.

Limitations

  • Standard back-adjustment embeds future action factors, causing look-ahead on price-level strategies
  • Only one end of an adjusted series equals prices that actually traded
  • Point-in-time adjustment is more expensive because the series must be re-derived per date
  • Feeds rarely document whether and how they adjusted, especially for dividends
  • Mixing adjusted and unadjusted prices across signalling and sizing creates silent inconsistencies

Common mistakes

  • Running fixed-rupee stops or absolute-level breakouts on a back-adjusted series and calling it realistic
  • Assuming a data feed is raw when it is silently back-adjusted, or double-adjusting an adjusted feed
  • Signalling on total-return adjusted prices but sizing positions on raw prices without reconciling them
  • Treating a dividend drop as a real loss because the series was not dividend-adjusted
  • Using an adjusted price as the actual fill price when the true traded price differed
  • Precomputing one adjusted series and using it for a strategy that needs point-in-time levels

Professional usage

Professionals store raw prices plus a dated corporate-action master and generate whatever adjusted view a task needs on demand, rather than trusting a single opaque adjusted feed. They use total-return adjusted series for performance and ratio-based signals, raw prices for execution and share-based sizing, and strict point-in-time adjustment whenever a rule depends on absolute price levels. They document the adjustment method so a backtest is reproducible, and they design strategies to be scale-invariant where possible so the adjustment choice cannot smuggle in look-ahead.

Key takeaways

  • Adjusted prices remove split and dividend jumps so returns and indicators behave across corporate actions
  • Standard back-adjustment rescales the past using later factors, embedding look-ahead for price-level rules
  • Point-in-time adjustment applies only factors known as of each date and avoids the leak
  • Use adjusted series for returns and indicators, raw prices for execution and share-based sizing
  • Store raw prices plus a dated action master so you control adjustment and can reproduce results

Frequently asked questions

What are adjusted prices?
Adjusted prices are historical prices rescaled to remove the mechanical jumps caused by splits, bonuses and dividends. The result is a continuous series on which returns, moving averages and other indicators are not corrupted by corporate-action discontinuities.
Why adjust prices at all?
Because raw prices jump for administrative reasons — a split divides the price, a dividend drops it — and those jumps are not real gains or losses. Adjustment makes each bar-to-bar change reflect an investor's true experience rather than a bookkeeping event.
What is back-adjustment?
Back-adjustment fixes the most recent price as the reference and rescales the entire earlier history to line up with it. Today's price stays equal to the real quote, but every historical price is then expressed in today's adjusted terms rather than what actually traded.
How can adjusted prices cause look-ahead bias?
A fully back-adjusted series uses the cumulative product of every split and dividend factor, including actions that happened after a given historical date. So a signal on an early bar is implicitly scaled by future actions, which is look-ahead for any rule that depends on price levels.
What is point-in-time adjustment?
Point-in-time adjustment applies, for each date, only the adjustment factors known as of that date, so early bars are unaffected by later actions. It reproduces what a trader would actually have seen at each moment and avoids embedding future information.
Does look-ahead from adjustment affect every strategy?
No. Scale-invariant logic such as percentage returns and ratio-based indicators usually gives the same decision on adjusted or raw data. The problem bites price-level logic — fixed-rupee stops, round-number breakouts, absolute channels — where the scaling changes what triggers.
Should I use adjusted or raw prices?
It depends on the task. Use adjusted, ideally total-return, prices for returns, indicators and comparison, and raw prices for execution modelling and share-based sizing, because raw prices are what a broker actually filled and what STT and brokerage applied to.
What is the difference between price and total-return adjustment?
Price adjustment handles only splits and bonuses, keeping a price-return series. Total-return adjustment also credits dividends back, so the series reflects the full outcome of holding the asset, which is what most benchmarks are quoted on.
How do I compute a split-adjustment factor?
For a split or bonus of ratio r-for-1, multiply prior prices by 1 divided by r; for a dividend, multiply by one minus the dividend over the previous close on the ex-date. Factors compound multiplicatively across multiple actions to give a cumulative factor.
Why do my adjusted and live prices differ?
Because a back-adjusted historical price is scaled to today's terms and does not equal what traded then. If you set a fixed-rupee level from an adjusted chart, it will not match what a live trader saw, which is why point-in-time data matters for absolute levels.
Is a free data feed adjusted or raw?
Often it is silently back-adjusted, and frequently it adjusts for splits and bonuses but not consistently for dividends. Because feeds rarely document their method, you should verify by checking a known corporate action before trusting the series.
Can I mix adjusted and raw prices in one backtest?
Only carefully. Signalling on total-return adjusted prices while sizing on raw prices is a common inconsistency that is hard to spot. Decide which series each part of the pipeline uses and reconcile them explicitly.
How do I store data to control adjustment myself?
Store raw traded prices together with a corporate-action master keyed by ex-date, then generate whatever adjusted view a task needs on the fly. This lets you reproduce, reverse or point-in-time-adjust the series instead of trusting an opaque feed.
Are adjusted prices related to survivorship bias?
They are separate issues but both concern historical faithfulness. Adjustment concerns whether prices are continuous and point-in-time correct; survivorship concerns whether the universe includes names that later delisted. A rigorous backtest handles both.

Voice search & related questions

Natural-language questions people ask about Adjusted Prices.

What are adjusted prices?
They are prices that have been smoothed out to remove the jumps from splits and dividends, so a stock's history looks continuous and returns make sense across those events.
How can adjusting prices cause a hidden bug?
The usual method rewrites the whole past using information about future splits and dividends. So your backtest can end up acting on knowledge it should not have had yet, especially with fixed price levels.
Should I use adjusted or raw prices in a backtest?
Use adjusted prices for returns and indicators, and raw prices for the actual fills and position sizes, because raw prices are what you would really have paid at your broker.
Why does my fixed stop behave oddly on adjusted data?
Because adjusted historical prices are scaled to today's level, not what traded then. A rupee stop set from an adjusted chart does not match what a live trader would have seen.
What is point-in-time price data?
It is price history adjusted the way it actually looked on each day, using only the splits and dividends known by then. It is the honest basis for any rule that cares about exact price levels.
Is a free price feed already adjusted?
Often it is, quietly, and it may adjust for splits but not dividends. Because feeds rarely say, check a known split or bonus in the data before you trust the series.

Sources & references

    Last reviewed 12 July 2026. Educational content only — not investment advice. Markets and rules change; verify current conventions with SEBI, NSE/BSE and your broker.

    ← Previous
    Missing Data
    Educational content only — not investment advice. Examples use illustrative numbers and simplified models. Backtested results are hypothetical and trading derivatives involves substantial risk. See our Risk Disclosure and SEBI Disclaimer.