The Chart Is a Set of Choices

Scale: what the vertical axis promises

A linear axis gives every dollar the same vertical distance. That makes it the right axis for a question about dollars held — the account, a position, a margin requirement — and the wrong axis for a question about performance over time, because a $1 move on a $5 stock is a 20% move and a $1 move on a $500 stock is 0.2%, yet they occupy the same square inch. On a long-term chart this flatters the recent past: a stock that went $20 → $100 → $80 has given back 20% of its value, and a linear chart draws that give-back as a quarter of the whole climb, because it is a quarter of the price range. A logarithmic axis gives every *percentage* the same vertical distance. That is the axis that matches how returns actually compound, and how a position actually behaves: a 10% move is a 10% move at $20 and at $500, because the equity curve is multiplicative. The rule of thumb worth carrying is short horizons, dollars matter, so linear is fine; long horizons and wide price ranges, percentages are the whole story, so use log. Every screenshot of a multi-decade chart that makes an old crash look catastrophic and an old bull market look flat has usually been drawn on the wrong axis for its argument. The practical consequence is not academic. A trendline drawn on a linear chart of a stock that has tripled is a curve in percentage terms, so the line that "held" five years ago sits in the wrong place relative to the return that matters. And because a linear axis compresses the early part of a compound move, breakouts and levels that were meaningful when the stock was smaller become invisible at the left of the chart — which is where the level a long-term holder cares about actually is. $10 → $100 → $50, read twice — Linear axis: The fall wipes out half the height of the climb — it reads as a collapse ← · Log axis: The climb spans one unit of height per doubling (10→20→40→80) and the fall is a single step down of the same size · Portfolio truth: A $10,000 position became $100,000 and is now $50,000 — a +400% outcome, not a collapse · When linear is right: Reading dollar risk: how much cash is at stake, and where the margin call sits The two axes agree at every point on the data. They disagree about which differences to make visible, and that is a choice with consequences rather than a matter of taste.

Adjustment and aggregation: the choices you cannot see

Prices are edited. When a stock splits 2-for-1, the historical prices are usually restated so the series is continuous — otherwise a $300 stock would appear to crash to $150 overnight and every indicator would scream. The same happens for dividends and spin-offs, and there are two conventions, not one: a fully adjusted series replaces old prices with restated ones, while a raw series leaves history as traded and shows the gap. Both are legitimate, and they differ in exactly the place a learner is most likely to be misled. On a raw chart a 3% dividend stock shows a small decline on every ex-date, and a stop placed under a level that predates a 4-for-1 split is sitting on a price the stock has never traded since. Aggregation is the second invisible choice: a bar is a container. A daily bar holds one session, a weekly bar holds five, and a 5-minute bar holds five minutes — so the four prices of the container, not the path inside it, are what you see. Two traders on the same stock with different containers will describe the same week differently and both be accurate: the weekly shows a clean uptrend while the daily inside it contains two 4% drawdowns and a gap. The container is chosen from the holding period, which is why the timeframe lesson later in this rung treats it as a risk decision, but the beginner point is simply that the container is a decision. The last two choices are visible but often ignored. Chart type: a line chart plots one price per period and quietly discards the intraday extremes, so it can show a serene trend through a week of violent swings; a bar or candle keeps the high and low, which is where the stops live. And point-and-figure abandons time altogether, drawing only when price moves by a fixed amount, which makes support and resistance crisp and makes duration invisible. None of these is better. A line chart that hides a 6% intraday range has hidden exactly the risk the position carries, and a P&F chart that shows a perfect base gives no clue whether that base took three weeks or three years. • Adjusted series: historical prices restated for splits and dividends — continuous, but not what traded. • Raw series: prices as they printed — honest gaps, and levels that can sit on prices no longer traded. • Aggregation: each bar is a container of four prices; the path inside it is discarded. • Line vs bar vs candle: line hides the extremes, where the stops are. • Point-and-figure: time disappears, so precision improves and duration information is lost. The single most common version of this mistake: a stop or a target set from a level on an unadjusted chart, then compared with a position whose cost basis was adjusted. The level and the basis are in two different units, and the arithmetic that follows is confidently wrong.

The horizontal axis is a choice too

Everything so far has been about the vertical axis, and the horizontal one is just as constructed. A standard chart draws one bar per calendar period, but that is one convention among several, and the choice decides which patterns can exist. Tick charts advance a new bar every N transactions, so busy stretches expand and quiet ones compress, which exposes short-term supply and demand at a resolution time cannot give. Range bars and Renko bricks advance only when price moves by a fixed amount, so a bar becomes a unit of movement rather than a unit of duration, which erases the empty hours and draws levels as clean horizontal bands. Volume bars do the same with traded size instead of transactions. The consequence is not cosmetic. A two-hour burst of selling inside a quiet session occupies one small bar on a time chart and disappears into the daily candle; on a tick or range chart the same burst is a wide, unmissable structure. Two traders examining the same afternoon can therefore reach different conclusions — and neither is wrong about the data, only about the container. The practical question is what the analysis is for: a thesis about where buyers and sellers are meeting is served by a movement-based axis, while a thesis about an event, a session or an earnings gap is served by time. The horizontal axis has a second and larger hidden version once the instrument is a future or a continuous series. To draw twenty years of a contract you must stitch dozens of individual months together, and the stitch is itself a decision: join them at expiry with the raw price gap intact, or back-adjust the whole history to remove the roll. The first shows the prices that actually traded and contains artificial jumps; the second is smooth and contains prices that never existed for the contract being drawn. Add the session choice — regular hours only, or the overnight tape as well — and the same crude-oil chart can be drawn four ways, each with its own trendline and its own gaps. What to carry out of this: the horizontal axis belongs to the same family of decisions as the vertical one, and it is the member most often left at the default. A movement-based axis answers where the market is trading; a time axis answers when it traded, and what events touched it. Before trusting a level or a breakout, know which of the two is on the screen — and when the instrument rolls or the session is trimmed, check whether the pattern survives the other convention. • Time bars: one per period. Best for event-driven reads — sessions, gaps, earnings, scheduled data. • Tick bars: one per N transactions. Best for seeing where order flow actually concentrated. • Range and Renko bars: one per fixed price move. Best for levels, because the empty time is removed. • Continuous futures: raw stitching shows real prices with artificial jumps; back-adjustment is smooth but historically fictional. • Session choice: regular hours only versus the full 24-hour tape, which changes every gap on the chart. The same move, four containers — The daily bar: One session, four prices; the intraday burst is inside it · A five-minute bar: The burst occupies one bar and the quiet afternoon occupies twenty · A 1,000-tick bar: The burst is a single wide bar; the quiet hours produce almost nothing ← · A $1.00 range bar: Only movement is drawn, so a level appears as a clean band None of these is more accurate. Each is a different answer to what a bar should measure — and the answer is what silently decides which patterns are visible.

Sources

• Linear versus logarithmic price scalesMurphy, “Technical Analysis of the Financial Markets”
• Splits, dividends and adjusted price seriesStandard data-vendor documentation on corporate-action adjustment
• Point-and-figure and time-independent chartingdu Plessis, “The Definitive Guide to Point and Figure”