Chart setup
Log or linear scale for trend lines?
Logarithmic — for anything that runs over years or moves more than about 30%. Not because it looks nicer, but because a straight line on a linear chart is secretly a claim that growth keeps slowing down. Here is the same trend line on Bitcoin, read by both rulers.
One trend line, two rulers
Bitcoin daily close, Jan 1, 2011 — Sep 25, 2026
The dashed line is drawn through two points and nothing else: the Dec 15, 2018 bottom at $3,182 and the Nov 21, 2022 bottom at $15,761. Same two anchors on both scales. Extended to today, they disagree completely.
The linear ruler says
$28,051
−66.6% off
The log ruler says
$75,256
−10.5% off
Bitcoin actually traded at
$84,100
Sep 25, 2026
Flip to linear and Bitcoin's history disappears
Bitcoin has run from $0.30 in 2011 to a peak of $124,724 — a 415,746× move. A linear axis has to fit all of that between one bottom and one top, so everything that happened at low prices gets pressed flat against the floor. That produces an inversion you can measure:
| Bull run | Move | Gain | Height on a linear chart |
|---|---|---|---|
| 2013 | $13.28 → $1,238 | +9,222% | 1.0% |
| 2017 | $785.99 → $19,424 | +2,371% | 14.9% |
| 2021 | $5,037 → $67,556 | +1,241% | 50.1% |
The 2013 rally gained 9,222% and occupies 1.0% of the chart. The 2021 rally gained 1,241% — 7× less — and occupies 50.1%. The biggest move looks the smallest. Every drawdown, every base, every trend line you would have drawn before 2021 is invisible on that scale.
Why a straight line only works on log
A trend line is not a drawing. It is a claim about a constant rate: this market has been moving at a steady pace and I expect that pace to hold. So the question is what "steady pace" means for a price.
Prices do not grow by adding a fixed amount, they grow by multiplying. An asset returning 10% a year does not add $10 every year — it adds $10, then $11, then $12.10. That is compounding, and it is how every market instrument behaves. Now look at what a straight line claims on each scale. On linear: price adds the same number of dollars every year — but as price rises, a fixed dollar amount is a shrinking percentage, so you have drawn a line that assumes growth is constantly decaying. On log: price grows the same percentage every year. A constant rate. Which is the claim you meant to make.
What to use, and when
| What you are doing | Scale |
|---|---|
| Trend lines on multi-year charts | Logarithmic |
| Anything that moved more than ~30% | Logarithmic |
| Comparing this cycle to previous cycles | Logarithmic |
| Comparing two assets at different prices | Logarithmic |
| Intraday or a few weeks, small range | Either — barely differs |
| Horizontal support and resistance | Either — a level is a level |
| Reading the dollar size of your P&L | Linear |
On TradingView the toggle is the small log button in the bottom-right corner, next to the price axis. It saves per layout, so a long-term layout and a scalping layout can each keep their own setting. Turn it on before you draw anything — redrawing every line because you started on the wrong scale is the annoying way to learn this.
Where log stops being enough
Log is less wrong, not perfect. Bitcoin's growth rate has been decaying cycle over cycle — each bull run returns less than the last — which is why the log projection above comes in −10.5% rather than exactly on target, while the linear one misses by −66.6%. Over a long enough arm even a log trend line drifts high.
The fix is not going back to linear, it is accounting for the decay — comparing cycles against each other rather than extending one straight line forever. That is what our Cycle Map does, and it is also why its bull-market comparison runs on a logarithmic axis: without it, the 2015 and 2018 cycles would be two flat lines along the bottom of the screen.
Frequently asked questions
Should I use a logarithmic or linear scale for trend lines?
Logarithmic, for anything that runs over years or moves more than roughly 30%. A trend line is a claim about a constant rate of change, and prices compound — they grow by multiplying, not by adding. On a log scale a straight line means a constant percentage per period, which is what a trend actually is. On a linear scale a straight line means a constant number of dollars per period, which quietly assumes the growth rate keeps slowing down. Drawn through Bitcoin's 2018-12-15 and 2022-11-21 cycle bottoms, the linear reading lands −66.6% away from where Bitcoin actually traded on 2026-09-25; the log reading lands −10.5% away.
What is the difference between a log and a linear chart?
On a linear chart every dollar of movement takes the same vertical space, so $10 to $20 looks identical in size to $1,000 to $1,010. On a logarithmic chart every percentage move takes the same vertical space, so $10 to $20 and $1,000 to $2,000 — both +100% — are drawn the same height. Linear measures absolute change; log measures relative change, which is what your return is actually made of.
Does switching to log scale change my trend line?
Yes, and this catches people out. The drawing stays anchored to the same two points, but the path between them and the extension beyond them both change, because 'straight' means something different on each scale. That means price can break a trend line on one scale while staying above it on the other. Pick a scale for a given type of analysis and keep it — otherwise you are comparing two different lines.
When is a linear scale still the right choice?
For short windows where the total range is small — intraday or a few weeks, under roughly 30% — the two scales are nearly identical, so it does not matter. Horizontal levels are also unaffected: support at $50,000 sits at $50,000 on both. Linear is also the honest choice when the absolute dollar amount is the point, such as reading the size of a position's profit and loss rather than its percentage return.
Is a log scale always more accurate for Bitcoin?
Less wrong, not perfect. Bitcoin's growth rate has been decaying across cycles — each bull run returns less than the last — so a log trend line drawn from the early years still points too high today. That is why the log projection here comes in −10.5% rather than exactly on target. Models that account for the decay, such as a power-law fit or cycle-relative comparison, are the next refinement. But that is a correction on top of log geometry, not a reason to go back to linear.
Bitcoin daily closes from 2011-01-01 to 2026-09-25, the same dataset that powers our Cycle Map. Trend-line anchors are the lowest daily close inside each cycle-bottom window; every figure on this page is computed from that data at build time, not typed in by hand.
By Menno van Ravels - 13 years in crypto, 3 bear markets survived, zero paid promotions