Direct answer
Logarithmic and linear chart scaling differ in what “the same distance on the chart” means. With linear scaling, a fixed chart distance corresponds to a fixed absolute price change. With logarithmic scaling, the same chart distance corresponds to a fixed proportional (percentage) change, which can make growth-like movements easier to compare across different price levels.
Advanced considerations are mostly about dependencies (what the chart is actually configured to do), edge cases (how scaling behaves under jumps, near-zero values, or mixed datasets), and implementation constraints (how you calculate returns and how your platform renders the scale). The key is to ensure the representation matches the question you are trying to answer and that you can verify it independently from the underlying price series.
Mechanism or definition
A linear price chart maps price values to a vertical axis so that equal vertical movement represents equal absolute changes in price.
A logarithmic price chart maps price values so that equal vertical movement represents equal proportional changes. In practice, this means the chart often looks more “even” when price is trending by percentages rather than by fixed increments.
What “proportional” means in calculations
To compare how scaling relates to changes, define two prices, an initial price (P_0) and a later price (P_1).
- Absolute change: (\Delta P = P_1 - P_0). Linear distance is tied to this concept.
- Proportional change: (r = P_1 / P_0). Logarithmic distance is tied to proportional movement, because logarithms convert ratios into differences.
A simplified mathematical intuition is that log transforms ratios: (\log(P_1) - \log(P_0) = \log(P_1/P_0)). That is why log scaling naturally corresponds to percentage-like change.
Where the dependency sits
Even if two people choose “log scale,” their results may differ because:
- The charting software may interpret the settings differently (for example, whether it logs raw price or an adjusted series).
- The axis may be configured to show log of price or another derived quantity.
- The analysis may still use linear measures (like absolute distances) while visually assuming log meaning.
So the advanced task is not just choosing the scale; it is aligning the scale with the computations and interpretation you perform.
Evidence or example
Consider the same proportional move from different starting prices.
- Example assumption: (P_0 = 100) and later price is (P_1 = 110) (a 10% increase).
- Another assumption: (P_0 = 200) and later price is (P_1 = 220) (also a 10% increase).
On a logarithmic chart, these two 10% increases correspond to the same proportional ratio (P_1/P_0 = 1.1), so the vertical change will be equal (given the same log transform rules).
On a linear chart, the absolute changes differ:
- First move: (110 - 100 = 10)
- Second move: (220 - 200 = 20) So the vertical distance is larger in the linear representation for the higher-priced series.
This illustrates the central comparison criterion:
- If the reasoning is fundamentally about percentage growth, log scaling often makes those proportional moves visually consistent.
- If the reasoning is about absolute thresholds (fixed pip distances, fixed nominal moves), linear scaling is more directly aligned.
Advanced implementation constraint: mixing measures
A common failure mode is mixing a log-scaled chart with an interpretation that assumes linear distances, such as measuring “equal distance” as if it meant equal absolute moves. Conversely, using log scaling while computing or thresholding with absolute differences can create mismatches between what you see and what you calculate.
To keep the analysis internally consistent, you can explicitly state which quantity you treat as primary:
- “I interpret distances as proportional changes.” (log-aligned)
- “I interpret distances as absolute changes.” (linear-aligned)
Limitations and risks
1) Near-zero and sign assumptions
Logarithmic transforms require positive inputs. If a price series includes non-positive values (rare for standard prices, but possible for derived series, spreads, or custom transformations), log scaling may be undefined or handled in a special way by the platform.
Even when prices are positive, data quality matters: missing points, outliers, or abnormal adjustments can create discontinuities that behave differently under log vs linear mapping.
2) Price gaps and event-driven jumps
If the series contains sudden jumps (for example, discontinuities from corporate actions in equities, or abrupt regime shifts in other markets), proportional change may produce visually large steps under log scaling, and absolute change may dominate under linear scaling. Both are “correct” transformations, but each can emphasize different aspects of the same history.
So the risk is not mathematical error; it is misreading what the chart emphasizes.
3) Changing the time window or reference point
Because log scaling emphasizes proportional change, the appearance of trends can change when you change the visible time range. A pattern that looks steady in log may look different in linear, or vice versa, especially across periods with large changes in the average price level.
This can lead to overconfidence if a viewer assumes the representation guarantees a stable interpretation across time.
4) Costs and execution are not “built into” the axis
Chart scaling does not incorporate trading costs, slippage, or execution constraints. Even if a representation helps visualize proportional movement, it does not ensure that the real-world outcome after costs matches the visual intuition.
Therefore, the limitation is: scaling helps interpret the price geometry, not the economic feasibility.
5) Jurisdiction and rule differences are outside the chart
If you are comparing tools or providers, regulatory and operational rules can differ by jurisdiction and platform. Chart scaling settings alone do not resolve those constraints.
Verification or next question
To independently verify that you are using the representation as intended, check these items:
- Confirm the chart setting: whether “log scale” truly applies to the raw price axis (or an adjusted series) and whether the platform uses log base (e), (10), or a normalized variant (base affects numeric scaling but not the relative geometry, as long as the transformation is consistent).
- Reproduce a simple transformation from two points: take (P_0) and (P_1), compute absolute change and proportional change (P_1/P_0), and verify which interpretation aligns with the chart’s visual distance.
- Test for consistency: apply the same method across multiple time windows to see whether your interpretation depends on a particular reference range.
A useful next question is: **What is your analysis question—absolute thresholds or proportional change?