What Is Slippage in Crypto? It Is Not a Speed Setting. It Is Your Price Floor.

Slippage is the gap between the price you were shown and the price you got. The tolerance box is not a speed control. It is the worst price you agree to accept before the trade cancels.

Updated 16 October 2026: the price-impact quotes were measured again and now sit beside the 13 August figures, the “nearly half of all sandwich attacks” figure is now credited to Jito’s analysis as Helius reports it, every quoted source is linked, and common questions are answered at the end.

That distinction is the whole article, and almost nothing written about slippage says it plainly.

People meet this setting for the first time in a hurry. A swap has failed once or twice, a box appears offering 0.5%, 1%, 5% or a field to type your own, and the obvious move looks like typing a bigger number. Bigger sounds like more room. More room sounds like the trade going through.

It does go through. That is exactly the problem.

The setting is a floor, not a dial

When you swap one token for another, you are trading against a shared pot that holds two things at once: a pile of the token and a pile of real money. Your trade changes the balance between them, and that balance is the price.

Between the moment you tap confirm and the moment the network actually processes it, other people are trading against the same pot. The price you were quoted is a photograph. The price you get is whatever the pot says when your turn arrives.

Uniswap, the exchange that made this style of trading popular, defines it in one line. Slippage, their documentation says, describes “alterations to a given price that could occur while a submitted transaction is pending.”

So the tolerance setting exists to put a limit on that. And here is the part almost every explainer leaves out: it is not a preference. It becomes a hard number written into your transaction, a floor for the least you will accept. The same Uniswap page is blunt about what happens underneath. If the trade “would return fewer output tokens than this value, the transaction reverts.”

Reverts means it does not happen. Your money stays where it is. You pay a tiny network fee for the attempt and nothing else moves. Nothing has been taken from you except that fee and a minute of your afternoon.

Read that as good news, because it is. A cancelled trade is the setting doing its job. It refused a deal you had already told it you would refuse. Out of every moving part in this process, it is the one piece of machinery working purely on your behalf, and the number you type is the only instruction it ever receives.

So what does 100% actually do?

It sets that floor at the ground.

A 1% tolerance says cancel this if I would end up more than 1% worse off than quoted. A 100% tolerance says do not cancel it at all. Any fill is acceptable. There is no longer a price bad enough to stop the trade.

Now, why do people do it? Not because they want a worse price. Helius, one of the companies that builds the plumbing Solana runs on, put the reason in writing in its research on how value gets taken out of ordinary trades. “Traders set high slippage,” it says, “not to accept worse prices but to ensure fast order execution.”

That is the misunderstanding, stated by people who watch it happen at scale. The setting has nothing to do with speed. It never did. It only ever answered one question, which is how bad a deal you would still take.

You can see how the confusion forms. Raising the number does make failed swaps stop, so it looks like a fix. What actually stopped was the cancelling, not the failing. The trades that used to be refused now complete, at the prices that were getting them refused.

How much room does a trade actually need?

This is answerable rather than theoretical, so this site measured it on $CATE, the token it covers: first on 13 August 2026, and again on 16 October 2026.

The quotes came from Jupiter, a Solana service that searches every pool trading a token for the best route for a swap. For each buy size, its quote service reports how far that one trade would push the price by itself, a figure usually called price impact. Nothing was actually bought. Most swap screens show the same kind of figure before you confirm.

In the 16 October quotes, a hundred dollar buy would move the price by under half a percent, and a thousand dollar buy by about the same. Even fifty thousand dollars, which is a great deal more than most people reading this will ever put into a meme coin, would move it about three percent. The August figures were a little higher at every size, but the scale is the same.

One detail matters here. Your own trade’s price impact is already inside the quote you are shown. As Uniswap puts it, slippage tolerances “establish a margin of change acceptable to the user beyond price impact.” So the tolerance only has to cover what other trades do to the pot in the moment before yours lands, and the chart gives a sense of scale for that. On those quotes, even a fifty thousand dollar buy landing just ahead of yours would move this price about three percent.

Now hold that next to a 100% setting. Typing 100 does not give you thirty times the safety. It gives away more than thirty times the room, and every bit of that gap is space where a worse price is still allowed.

Nothing forces the pot to fill you at the best available price. It fills you at whatever price the conditions produce, and your floor is the only thing that ever says no.

Something lives in that gap

Here is why the empty space is not merely unused.

There are automated programs that watch trades waiting to be processed. When one of them spots a trade with a lot of room underneath it, it can buy just ahead of you, let your trade push the price up further, and sell straight after. You get filled at the top of the move it created. It keeps the difference.

The nickname for this is a sandwich, because your trade ends up in the middle of two others.

The scale of this is not small, and it is documented rather than rumoured. In its report on the subject, Helius traced a single one of these programs, known as Vpe, over thirty days from 7 December 2024 to 5 January 2025. It ran 1.55 million of these sandwiches and took 65,880 SOL, worth about $13.43 million at the time. The average haul per trade was about $8.67, which points to ordinary-sized trades rather than whales.

The same report cites internal analysis from Jito, a Solana infrastructure company, which attributes nearly half of all sandwich attacks on Solana to that one program.

Two details from that report land close to home for anyone reading this site. Helius names the vulnerable group directly: meme coin traders, it writes, “tend to set high slippage tolerances when trading in illiquid, highly volatile assets.” And of the twenty most-sandwiched tokens it identified, sixteen had been launched on pump.fun. That is the same launchpad $CATE came from, and the launch mechanics are pulled apart in what a pump.fun bonding curve actually is.

None of this means every wide trade gets taken. It means the room you leave is a standing offer, and there is an industry set up to accept it.

Why small pots make all of this worse

Everything above scales with how much real money is sitting in the pot you are trading against.

A deep pot absorbs your trade without noticing. A shallow one moves hard, so the same purchase arrives worth noticeably less than the screen promised, and a wide tolerance lets it happen quietly. This is the practical consequence of something measured in how meme coins work, where the headline market cap turned out to be many times larger than the money actually available to trade against.

It is also why the depth figure is worth looking at before the price. Where to find it, and which numbers on those sites quietly mislead you, is how to read a DexScreener chart. The live figures for $CATE are always on this site’s stats page.

What to take from this

This post will not tell you what number to type. That depends on the token, the size of your trade and the day, and anyone handing out a single figure is guessing on your behalf.

What is worth carrying away is smaller. The box is not asking how badly you want the trade. It is asking what you would still accept, and that is the question to answer.

A trade that keeps failing is telling you something real about the pot. It is information, not an obstacle to be overpowered with a bigger number. Sometimes the answer is a little more room. Sometimes it is that the thing you are buying cannot absorb what you are trying to spend. Selling runs into the same wall from the other side, which is covered in how to sell a meme coin.

If you want to see the other end of this, where people lose everything rather than a few percent, that is how wallets actually get drained. And the setting is one line of a longer process covered in how to buy $CATE.

Frequently asked questions about slippage