Where Crypto Prices Come From: There Is No Single Price

Why every site shows a different Bitcoin price: order books, spreads, volume-weighted averages, and why your exchange never matches the number you saw.

The price that made my friend think he was scammed

My admission for this post involves someone else's money, which makes it worse. A friend asked me what Bitcoin cost, I read him the number from a price site, and he went off to buy some on an exchange. He paid noticeably more than the number I quoted, once the exchange's own market price, the spread, and fees all landed, and he came back convinced the platform had cheated him. It had not. I had. I quoted a global reference number as if it were a price anyone could actually transact at, because I did not yet understand that there is no such thing as the price of Bitcoin.

There are thousands of prices, one per trading venue per moment, and the tidy number on any tracking site is a computed summary of them. Once you see how that summary is built, every confusing discrepancy between sites, apps, and your own exchange becomes not just explicable but predictable.

A price is just the last trade somewhere

Unlike a stock with its primary listing, Bitcoin trades simultaneously on hundreds of independent exchanges, each running its own isolated order book. An order book is two lists: bids, what buyers currently offer to pay, and asks, what sellers currently demand. When a bid and an ask meet, a trade executes, and that trade's value becomes the exchange's last price until the next trade replaces it.

So when an exchange displays a price, it is reporting the most recent trade in its own book, nothing more. Two exchanges with different customers, different fees, and different local currencies will mechanically produce different last prices at the same instant. Usually the gap is small, because arbitrage traders profit by buying on the cheaper venue and selling on the dearer one until the gap closes. But small is not zero, and in volatile minutes or on thin markets the gaps widen exactly when you care most.

The spread: the price is actually two prices

Even on a single exchange, the price you can buy at and the price you can sell at are never the same number. The gap between the best ask and the best bid is the spread, and it is the first cost my friend hit. Worked example with clean numbers: if the best bid is 60,000 dollars and the best ask is 60,060, the spread is 60 dollars, which is 0.1 percent of 60,000. Buy instantly and you pay the ask; sell instantly and you receive the bid. The midpoint that price sites often show is a number at which literally nobody is trading.

Spreads scale with liquidity. Major pairs on major exchanges stay tight; smaller coins on smaller venues can carry spreads of a full percent or more before any fee is counted. Larger orders also walk deeper into the book, filling against progressively worse prices, which is called slippage. None of this is hidden, but none of it appears in the headline number either.

How aggregators build the number you actually see

Tracking sites and converters do not pick one exchange; they compute a volume-weighted average across many, so venues where more real trading happens count for more. Tiny worked example: suppose exchange A trades 100 coins at an average of 60,100 dollars while exchange B trades 300 coins at 60,020. The volume-weighted price is 100 times 60,100 plus 300 times 60,020, divided by 400 total coins: 24,016,000 divided by 400, which is 60,040. B's price dominates because three quarters of the actual volume happened there.

Real methodologies, like CoinGecko's linked below, layer safeguards on top: excluding venues with suspect volume, dropping stale or outlier ticks, and converting everything into a common currency. That last step matters more than people expect, since a coin trading against dollars, euros, and stablecoins has its aggregate price shaped by currency conversion rates too.

The practical consequence: two aggregators with different exchange lists and different weighting windows will legitimately disagree by small amounts, and both will differ from any single exchange. When you use a crypto price converter, you are getting exactly this kind of computed reference rate, ideal for valuing holdings and comparing moments in time, and never a guaranteed execution price.

Why your exchange never matches the tracker

Stack the pieces and the mystery dissolves. The tracker shows a volume-weighted global average of last trades. Your exchange shows its own book's last trade, which sits somewhere inside the global cloud of prices. You then transact not at that last price but at the current bid or ask, cross the spread, possibly slip deeper into the book, and pay a trading fee on top. Each step is small; together they routinely explain a difference of a percent or more on modest trades, and more on thin pairs.

This layering is also why unit math confuses newcomers. Prices are quoted per whole coin, but nobody is obligated to buy whole coins. At 60,000 dollars per Bitcoin, 20 dollars buys a three-thousandth of one, which is easier to reason about in satoshis, the hundred-millionth subdivisions of a Bitcoin: 20 divided by 60,000 is 0.000333 coins, or 33,333 satoshis, a conversion a satoshi converter does without the zeros tripping you.

Using reference prices without being used by them

Reference prices are excellent for what they are: valuing a portfolio, comparing today with last month, setting alerts, and doing planning math. Position sizing is the best example of legitimate use: deciding how much to buy based on account size and acceptable loss. With a 5,000 dollar account risking 1 percent, 50 dollars, on a trade entered at 60,000 with a stop at 58,800, the risk per coin is 1,200 dollars, so the position is 50 divided by 1,200, about 0.0417 coins, roughly 2,500 dollars of exposure. A position size calculator runs that arithmetic in seconds, and the reference price is plenty accurate for it.

Where reference numbers mislead is execution and emotion. For execution, the only price that matters is the book in front of you at the moment you click. For emotion, remember that aggregated numbers move with sentiment cycles that are themselves measurable; a glance at a crypto fear and greed index alongside the price chart is a useful reminder that the number you are staring at is partly a mood ring.

So the honest summary: there is no single price, only a well-built average of many. Quote it, plan with it, track it. Just never promise it to a friend as the number he will pay. I only needed to make that mistake once.

Questions people ask

Why do two price sites show different Bitcoin prices at the same time?

Each aggregates a different set of exchanges over slightly different time windows with its own weighting and outlier rules. Both numbers are honest summaries of a market that has no single price, so small disagreements are expected, not evidence of an error.

Why did I pay more than the price shown when I bought?

You crossed the spread by paying the ask rather than the midpoint or last price, may have slipped deeper into the order book on a larger order, and paid the exchange fee. Each cost is individually small, but they stack on top of the reference number you saw.

What is a volume-weighted average price?

An average where each exchange's price counts in proportion to how much trading actually happened there. If 300 coins traded at 60,020 and 100 traded at 60,100, the weighted average is 60,040, sitting closer to the venue with three times the volume.

Do I have to buy a whole Bitcoin?

No. Bitcoin subdivides to a hundred-millionth, called a satoshi, and exchanges sell fractional amounts routinely. At 60,000 dollars per coin, 20 dollars buys 33,333 satoshis. The per-coin quote is a unit price, not a minimum order.

Which price should I use for taxes or portfolio tracking?

A consistent reference source, applied the same way every time. Tax authorities in the US and UK generally accept a reasonable, consistently used valuation method. Consistency matters more than which reputable aggregator you pick; switching sources between entries is what creates problems.

Read next

All articles