Round Numbers 00 and 50: Why Orders Cluster There and Where to Put Your Stop

A round number is a price ending in 00 or 50, such as 1.1000 on EURUSD or 4,000 on gold. Many traders believe price “respects” these levels, and there is research that explains a real mechanism behind it: take-profit and stop-loss orders are not spread evenly, they pile up around round numbers in two different ways. This article explains that research carefully, its limits, and how to place your own stop so it isn’t sitting in the middle of everyone else’s.

Why people like round numbers

Shops price things at 99 instead of 100 because people read numbers from the left and 99 feels much cheaper. That is psychological pricing on the seller’s side.

Markets show the other side of the same habit. When people have to choose a price for an order, they pick numbers that are easy to think about, remember and say. “Take profit at 1.1000” sounds more natural than “take profit at 1.1003”, even though there is no economic reason behind it. When many people think the same way, their orders end up at the same price.

Carol Osler’s research

Carol Osler, an economist who worked at the Federal Reserve Bank of New York, wrote two widely cited papers on this.

• 2000, in the New York Fed’s Economic Policy Review. She tested the support and resistance levels that financial firms sent to their customers each day against randomly chosen levels. Intraday trends stopped at the firms’ levels more often than at the random ones. So support and resistance carried some real information, but the paper did not say why.
• 2003, in the Journal of Finance. She went looking for the reason, using the actual stop-loss and take-profit orders that customers had placed with a large bank in the currency market.

Without the numbers, the 2003 findings come down to this:

1. Take-profit orders pile up exactly at round numbers, especially prices ending in 00
2. Stop-loss orders cluster just beyond round numbers. Buy stops sit just above, sell stops just below
3. Together these fit two things technical traders have said for a long time: trends tend to stall or reverse at round numbers, and once price gets through a round number, the move tends to speed up

The mechanism: why piled-up orders make price stall or speed up

You need to separate two kinds of order.

• The take profit of a buy trade is a sell order waiting in the market. It works like a limit order. When price rises to 1.1000, a large amount of selling is waiting there, and the buying pushing price up has to absorb all of it first. So price often stalls
• The stop loss of a sell trade is a buy order waiting above 1.1000. It works like a stop order: once touched, it becomes a buy at the market price. That buying pushes price up into the next stops, and so on (a stop cascade). So price often speeds up

Put simply, take profits absorb a move and stops amplify it. More on how each order type behaves in Order types explained, and on liquidity in the currency market in FX market structure and liquidity.

The chart below comes from a simulation. It shows what a histogram of orders looks like if people place them with these habits.

This is the same idea the ICT school calls a liquidity sweep or stop hunt: price runs past an obvious level to collect the stops, then comes back. Tests of that idea are in ICT tested on gold and Smart Money Concepts tested.

Limits to know before you rely on it

• The data is old and from one place. The 2003 paper used one bank’s orders from the late 1990s to 2000. Today’s market has far more algorithms and electronic platforms
• It is a statistical tendency, not a rule. Price stalling at round numbers more often than at ordinary levels does not mean it stalls every time, or tell you for how long
• You cannot see the real order book. Spot FX has no single central exchange. Your chart does not show where orders are waiting or how many
• Everyone knows about it now. If everyone knows stops sit below 1.1000, so do the people holding large positions, which is why runs to collect stops and snap back are common

Our own tests did not measure round numbers directly. But in our backtests on gold M1, support and resistance that “lined up” with Fibonacci levels happened about as often as with made-up placebo levels (details and conditions in Support and resistance, supply and demand). The same lesson applies here. The right question is “does price stall at 1.1000 more often than at 1.1037?”, not “has price ever stalled at 1.1000?”. Backtest results do not predict future results.

A round number is not a wall. It is a place where many other people have orders. Use it to know where price may change pace, not to guess direction

Where to put your own stop loss

If your stop sits in the same pile as everyone else’s, you face two problems. It gets hit easily, because price often runs just past the round number to collect that pile. And when many stops fire at once, you may get filled at a worse price than you set (slippage).

Simple rules that work:

1. Start from price structure. Your stop belongs where “if price gets here, my idea is wrong”. Only then check whether a round number is nearby
2. Don’t put it exactly on the round number, and don’t put it just 3–5 pips past out of habit, because that is exactly where everyone else’s orders are
3. If a round number sits between your entry and your stop, put the stop a further distance beyond it based on ATR, not a fixed number of pips. For example 0.5–1 times ATR on the timeframe you trade. That range is a starting point to adjust, not a proven best value
4. Size the lot from the full distance. A farther stop needs a smaller lot, so the money at risk stays the same

The buffer helps you avoid being taken out by a wick that runs to collect stops and comes straight back. It does not protect you from a real break. If price breaks through the round number and keeps going, the buffered stop gets hit too.

Worked example: a stop near 1.1000, and the lot size

Say your balance is 5,000 dollars and you risk 1% = 50 dollars per trade. You buy EURUSD at 1.1032, with the idea that price will hold above 1.1000. ATR(14) on your chart = 11 pips (an assumed value). One pip on 1 lot of EURUSD = 10 dollars (what most brokers use; check your own contract specification).

Lot = money at risk ÷ (stop distance in pips × 10), rounded down to 0.01.

Option	Stop loss	Distance	Calculated lot	Lot opened	Actual loss if hit
A exactly at the round number	1.1000	32 pips	50 ÷ 320 = 0.156	0.15	0.15 × 32 × 10 = 48.0
B 5 pips below it	1.0995	37 pips	50 ÷ 370 = 0.135	0.13	0.13 × 37 × 10 = 48.1
C 1 ATR beyond it	1.0989	43 pips	50 ÷ 430 = 0.116	0.11	0.11 × 43 × 10 = 47.3

Option C is 1.1000 − 11 pips = 1.0989. Its distance from the 1.1032 entry is 1.1032 − 1.0989 = 0.0043 = 43 pips.

All three options risk almost the same, no more than 50 dollars. The difference is location. A and B sit right where price tends to run to collect stops. C sits beyond that pile, in exchange for a smaller lot, 0.11 instead of 0.15. If price reaches the same target, C makes proportionally less. Try your own numbers with the lot size calculator, and read the method in Position size from your stop loss.

One more thing: 1.0989 is just below 1.0990, which also ends in 0. Any pile there is likely much smaller than at 00, but if you want to be clear of it, move out another 1–2 pips and recalculate the lot.

Take profit near a round number

The same mechanism works in reverse. If the target of a buy is exactly at 1.1100 or just past it, you are placing your order where everyone else’s is, and that is where price tends to stall. Some traders therefore set their take profit a little before the round number, for example 1.1095. The reasoning follows from the mechanism above, but we have no test showing how much it helps, and a closer target changes your RR. Always check it against the break-even win rate.

Round numbers on gold: fixed distances don’t carry across years

Gold’s round levels are every 100 or 50 dollars, but what that means changes with the price.

• At 1,800 dollars, a 100-dollar step = 100 ÷ 1,800 = 5.56% of the price
• At 4,000 dollars, a 100-dollar step = 100 ÷ 4,000 = 2.50% of the price

The same thing happens to a stop “with a 5-dollar buffer”. In our measurements on gold (XAUUSD) 1-minute bars, the median size of price swings, measured the same way every year, was about 13.6 dollars in 2022 and about 49.6 dollars in 2026 (data up to August), 3.7 times bigger. Measured as a share of price, it only went from 0.76% to 1.18%. A buffer set in fixed dollars or pips is therefore a different size each year. A buffer based on ATR or a percentage of price adjusts by itself. Backtest results do not predict future results.

Bitcoin has heavily discussed round numbers too, such as 100,000 dollars. The same idea applies, but crypto trades across many venues and is very volatile, so the buffer has to be wider as well.

A good stop sits where your idea is proven wrong, not where everyone else puts theirs. If it has to go past a round number, add an ATR buffer and cut the lot

Common mistakes

• Putting the stop at 1.0995 because “I already left room”. 5 pips below the round number is the middle of everyone else’s sell stops. It gets hit easily and may slip
• Buying because price “reached the round number, so it must bounce”. The research only shows a tendency to stall. It says nothing about how far a reversal goes, and if price breaks through, it often moves fast
• Widening the stop without cutting the lot. A 43-pip stop at 0.15 lot loses 64.5 dollars, not 50. More in Wide or tight stop loss
• Using a fixed pip buffer for every market and every year. 5 pips on a quiet day and on a news day are not the same, and neither is gold at 1,800 versus 4,000
• Trusting what you see on past charts. You can always find a bounce exactly at a round number to show off. Count every time price passed the level, including the times it didn’t bounce. Read Why charts look easy in hindsight