How to Read a Price Chart
Without Fooling Yourself

A price chart is the fastest way to become confidently wrong about a card. It takes a year of separate, awkward, badly-labelled transactions and turns them into one smooth line, and a smooth line reads as a fact. It is a summary, and whoever drew it settled five questions first: what counts as a sale, how to average them, where to start the vertical axis, how much history to show, and what to draw for the weeks when almost nothing traded. All five are recoverable in a minute, and they change the reading completely.

Find out what one point is made of

THREE CHARTS THAT LOOK IDENTICAL

Every chart you will meet is one of three things, and they are drawn with the same line in the same colour. A sale-based chart plots what people paid. An average-based chart plots a rolling calculation over recent sales, so it is smoother, slower, and lags the thing you are trying to catch. A listing-based chart plots what sellers were asking on a given day, which — as the chapter this article sits in keeps saying — is a number with one signature on it.

You cannot tell which is which by looking. You can tell in ten seconds by reading the label above the chart and, failing that, its help page. Bridge the vocabulary once and you are done with it.

Smoothing is not a flaw — a rolling average is exactly what you want for a card that trades daily. But it has a cost: the chart shows a gentle slope where the market had a cliff, because the average is still carrying last month's sales down the hill.

Look at where the axis starts

THE CROP DOES THE TALKING

Almost every price chart on almost every site crops the vertical axis to fit the data. This is a reasonable default and it is also the single most effective way to make a quiet card look dramatic. Below are the same four monthly figures — $48, $47, $49 and $58 — drawn twice. Nothing about the card changed between the two panels. Only the bottom of the axis did.

The habit is small: before you read the shape, read the bottom-left number. If the axis does not start at zero, convert the move to a percentage before you react to it. $49 to $58 is a real rise — about eighteen per cent — and worth knowing about. It is not the vertical wall the left panel draws, and that difference is the difference between buying carefully and buying in a hurry.

A chart cannot show you a sale that nobody made. It draws a line through the gap anyway.

Widen the window before you decide anything

TWELVE MONTHS, NOT NINETY DAYS

Most sites open on a three-month view, because it loads quickly and looks tidy. Three months is long enough to feel like history and short enough to miss a reprint, a rotation or a ban list. Push the window to a year and most charts stop being ambiguous. Here is the same card over twelve months, with the number of sold listings under each point.

On a ninety-day view, the last three points are all you get: $47, $49, then $58. That is a chart of a card going up, and nothing in it tells you the card lost more than a third of its value earlier in the same year. On the twelve-month view the story is obvious, and it is the opposite story.

Read the row underneath the line

HOW MANY SALES IS THIS POINT?

Volume is the part of a price chart people never look at, and it is the part that decides whether a point deserves a reaction. Every month on that chart is drawn with the same weight of line. They are not the same quality of evidence. The reprint month rests on twenty-two sales. The current month rests on two.

That single fact changes which number you would actually quote. There are three defensible readings of this chart and they are a long way apart.

The twelve-month median is the interesting failure here. It is calculated correctly, it uses every sale, and it is useless — because averaging across a structural break averages two different objects. Once a chart contains a step, the only honest window starts after the step.

Under about five sales in a month, treat the point as a data point and not as a price. Two sales can be one enthusiast and one mistake.

Tell a step from a spike

ONE REVERTS, THE OTHER DOES NOT

Nearly every shape worth noticing on a price chart is one of two things, and the test is simple enough to run from the chart alone.

• A step moves the level and stays there. It comes from something that changed the supply or the legality of the card — a reprint, a rotation, a ban, a set entering a new format. It usually arrives with a volume surge, because everybody trades at once, and the new level holds after the volume falls back.
• A spike moves the price and comes back. It comes from attention: a tournament result, a video, one dealer buying out a listing page. Volume rises with it and the price returns as the attention moves on, often inside three weeks.

So: did the volume stay elevated after the move, and did the price settle at a new level or drift back to the old one? On the chart above, the drop at month −6 is a step — twenty-two sales, then seventeen, and six months that never go back above $52. The rise at the right edge fails both tests: no volume behind it, and no time to hold anything.

Which does not make it wrong. It makes it undecided. A spike and the first month of a step look identical from inside the first month, and the only thing that separates them is what happens next.

Write the reading down with its window

THE LAST TEN SECONDS

A figure taken off a chart is worth keeping only if you keep the frame it came from with it. Write a sentence, not a number: median $49 over the six months since the reprint, 83 sales, plain print, raw. That survives being read back in six weeks, because it says which window it covers and why the window starts where it does.

And it makes the next conversation shorter. When somebody tells you the card is climbing, you are not disagreeing with them — you are asking how many months they are looking at, and how many sales are under the last point. Most of the time, the honest answer to the second question is two.