Stop loss placement: invalidation, noise and the size trade-off

Start in plain English

A stop loss is the exit you plan for the case where the trade idea is wrong. This lesson is about where it goes. How stop orders and stop-limit orders work is covered in US Day Trading Fundamentals: Stop and stop-limit orders.

The core idea: a stop marks the price at which the reason for the trade no longer holds. That is a statement about the chart, not about how much loss feels acceptable. Once the stop is placed, its distance from entry sets the share count (at a fixed dollar risk), and it sets how often ordinary back-and-forth movement, called noise, reaches it before the idea has had a chance to work. A tight stop allows more shares and gets hit by noise more often. A wide stop survives noise and allows fewer shares. Neither is better by default.

Two paths from the same start

All bars, prices, sizes and costs in this example are Illustrative. Calder Wren Systems is fictional.

The setup at 10:08 ET. Calder Wren pushed from $5.80 to $6.30 between 09:58 and 10:01, pulled back to a low of $6.02 at 10:04, and closed the 10:07 bar at $6.17. The idea: the pullback has held, and the push may continue. The idea is wrong if price breaks the pullback low.

Two stop plans for the same entry, each risking $90 (1R) and filled at $6.18, with $0.005 per share in round-trip fees and a time stop at 10:18:

To keep the comparison simple, the share counts use the stop distance alone: $90 ÷ $0.06 = 1,500, and $90 ÷ $0.19 ≈ 473, rounded down to 470. Fees and slippage are left out of the sizing but included in every result below, which is one reason a stopped trade loses more than 1R. The full rule from Position sizing adds them to the per-share risk and sizes smaller: with the fee alone, $90 ÷ $0.065 ≈ 1,384 and $90 ÷ $0.195 ≈ 461 shares.

The review note: "Same entry, same idea, same dollar risk; outcome decided by which stop the path reached first." These three paths happen to favour the wider stop. A fourth path favours the tight one: price runs straight to $6.30 without dipping below $6.12, the tight plan earns +1.92R on 1,500 shares, and the wide plan holds 470 shares to the time stop at $6.36 for +0.91R. The rest of this lesson covers how to tell invalidation from pain, how to measure noise, hard and mental stops, halts and time stops, and how to compare two stop rules on your own log.

Invalidation vs. pain threshold

Two ways to choose a stop:

• Invalidation: "If price trades below $6.02, the pullback-held idea is wrong." The stop goes just beyond that level, with room for spread and a few cents of slippage.
• Pain threshold: "I don't want to lose more than 6 cents a share." The stop goes wherever that number lands on the chart, which may be in the middle of ordinary movement.

A pain stop answers a sizing question with a price. Sizing is better handled by share count: choose the stop from the chart, then size so that the stop costs your planned dollar risk (Position sizing). If the invalidation level is so far away that the size becomes tiny or the reward is under 1R, the honest answer may be no trade.

Measuring noise: spread and the typical 1-minute range

Noise is the movement a stock makes when nothing about your idea has changed. Two simple measures:

• Spread. A stop two cents below entry in a stock with a two-cent spread can be reached by the quote moving one tick.
• Typical 1-minute range. The average of high minus low over recent bars. In the Calder Wren example the last five bars averaged 8 cents. A stop at 0.75 times that range sits inside one ordinary bar; a stop at 2.4 times it needs a larger move.

These ratios are descriptions, not rules. Writing them down for every trade lets your log answer the question this lesson cannot: how often stops at a given ratio were hit before the trade reached its target, in your stocks.

Hard stops and mental stops

• Hard stop: a live stop order. If it is held by the broker or a venue rather than by your own platform, it triggers even if your platform freezes or your internet drops. Depending on the broker, it is held on the broker's servers or sent to a venue; either way, once triggered it fills at the next available price.
• Mental stop: a price you intend to act on. It needs you to see the level and press the key. Under stress, mental stops tend to move.

For scalps that last seconds to minutes, a hard stop placed right after the fill removes one decision from the worst moment. The stop's quantity should match the filled shares; the DAS course lesson Orders, fills and protection are separate states shows why in one platform.

Stops into halts

A stop cannot execute during a trading halt or LULD pause. If a stock reopens below your stop, the stop fills there. In Path B, a pause and a reopening at $5.70 would have filled the wide stop near $5.68: ($5.68 − $6.18 − $0.005) × 470 = −$237.35, or −2.64R (Illustrative). No placement fixes this; only size limits it. See trading halts, LULD and What happens to your orders during a halt in US Day Trading Fundamentals. A FennEdge data study is planned to measure reopening prices against halt prices.

Time stops

A time stop exits if the trade has not reached target or stop by a set time. It addresses a third outcome that price stops ignore: nothing happens. In Path C, the time stop turned a chop into a −0.08R scratch instead of an open position drifting toward the stop.

Observe vs. infer

At 10:08 ET in the Calder Wren example (fictional; all figures Illustrative):

Odds box

The two rules feel very different: the tight one loses more than half the time, the wide one spends more trades going nowhere. On these assumed numbers the difference in expectancy is a few hundredths of an R, far smaller than the uncertainty in any small log. The point is not that one rule wins. It is that the stop rule changes the whole shape of the outcome distribution, so it needs to be fixed before you count.

What shifts the odds

These are hypotheses to record, not findings:

• Stop distance as a multiple of the recent 1-minute range.
• Spread in cents and as a share of the stop distance.
• Distance to a round number ($6.00, $5.50) and to obvious prior lows.
• Time of day: early-session bars tend to be wider. FennEdge's Gapper HOD Timeline shows when gappers printed their highs (guests, Free and Plus see the latest completed session; earlier sessions need Pro or Max, as of 2026-10-05), which is context for how volatile the first hour of those days was.
• Recent halts in the same stock.
• Volume trend through the pullback.

Cost of being wrong

Calder Wren, $90 risk per plan (Illustrative):

The tight plan's 1,500 shares also carry more halt risk if the pause comes before its stop: the same reopening at $5.70 would cost ($5.68 − $6.18 − $0.005) × 1,500 = −$757.50, or −8.42R.

How many trades before this means anything

Comparing two stop rules means comparing two outcome distributions, each with a wide interval. With the Illustrative splits above, the per-trade spread of results is about 1.3R to 1.5R; after 50 trades per rule, a rough 95% range for each rule's average is about ±0.4R (normal approximation, computed from the assumed splits). The difference between +0.07R and +0.10R is invisible at that sample size. Log the stop distance, the recent-range multiple and the path for every trade, and read How many trades before your numbers mean anything.

What this lesson does not claim

• It does not claim tight or wide stops are better. Every split above is Illustrative.
• It does not claim that prior lows or round numbers act as support, or that stops cluster at any level.
• A hard stop limits the trigger, not the fill; halts and thin books can push the fill beyond it.
• Calder Wren Systems and all prices are fictional.

Go deeper

• FennEdge first.
  • In-Play Days shows 1-minute charts of past active days, useful for measuring typical 1-minute ranges in the stocks you trade. Without Max, including as a guest, you can open each ticker's most recent active day in full; earlier days and the two-year history need Max (as of 2026-10-05).
  • Gapper HOD Timeline for time-of-day context on gappers.
  • DAS course: Orders, fills and protection are separate states.
• Stop clustering. Research on currency markets (Osler, 2003) found stop orders concentrated just beyond round numbers. Whether this applies to small-cap stocks is a separate question; treat it as a hypothesis.
• Volatility-based stops. Stops set at a multiple of recent range (as above) widen automatically in fast stocks and shrink the share count with them.
• Trailing stops move with price after entry. They change both the exit and the path shares; count them as a separate rule.
• Related lessons: Risk/reward and breakeven win rate, Slippage, A trading plan as a hypothesis, VWAP trading strategy.

Exercises

Six exercises follow this lesson: three concept checks, one calculation (expectancy of the tight-stop rule after slippage), one replay of the Calder Wren bars that asks you to separate what the screen showed from inference, and one scenario about an invalidation level that sits far from entry. Exercises, answers and progress are available after signing in.

Primary sources

• FINRA. Rule 5350: Stop orders.
• Osler, C. L. (2003). Currency orders and exchange rate dynamics: An explanation for the predictive success of technical analysis. The Journal of Finance 58(5).
• LULD Plan. Limit Up-Limit Down Plan.

Sources checked 2026-10-05. Calder Wren Systems and all bars, prices, outcome splits and costs in this lesson are fictional and Illustrative. This lesson is education, not a recommendation to buy or sell any security. Past patterns do not indicate future results.