Order types for entries

Monday, 10:12 ET (22:12 ICT), a synthetic NQ session. Your plan says: buy the first pullback to 29,500.00 after this morning’s breakout. You rest a buy limit at 29,500.00. NQ trades down to 29,500.25, one tick above your price, and rallies 20 points without you. Ten minutes later another trader’s buy-stop at 29,516.00 is triggered by a jump and fills at 29,519.00, three points above the trigger. Each order did exactly what it promises. This lesson is about choosing that promise on purpose for each entry family, and about the uncomfortable fact behind every limit: it fills most easily when the market is moving through it.

Key terms

Limit order (lệnh giới hạn): an order to buy at your price or lower, or sell at your price or higher Definition. It promises the price, not a fill.

Stop order (lệnh dừng, stop-market): an order that rests until price trades at the trigger, then becomes a market order Definition. It makes a fill very likely, not the price. On CME Globex, market and stop orders carry a protection limit, and a price band or a pause can stop matching, so even a stop can fill late or in part (M0.03).

Stop-limit order (lệnh dừng giới hạn): a stop whose trigger sends a limit order instead of a market order Definition. It caps the price; if the market jumps past the limit, you get no fill.

Adverse selection (lựa chọn bất lợi): a resting order is filled by whoever chooses to trade against it, and the traders most eager to do so are often the ones who know or move more than you Definition. Markout (độ lệch sau khớp) measures it: the mid-price a few seconds after your fill minus your fill price, signed so that a negative number means the market moved against you Definition.

Chart first: three orders, one move

How it works

Match the order to the entry family

Each entry family from M6.01 asks the order for something different. The pairing below is a house rule House rule, not a measured edge.

• Pullback and retest entries (M6.03): a limit at the price written in the plan. The limit protects the price you planned. It does not make that price special: prior-day levels did not beat randomly placed levels in our lab’s tests (M6.03), so a missed fill is a missed trade, not a missed edge.
• Breakout and continuation entries (M6.02, M6.05): a stop at the trigger price. You pay for participation with slippage, which you log in ticks and compare with your plan (M10.02).
• Rejection entries (M6.04): the order sits at the trigger (the close back inside the level), never at the extreme. A limit at the extreme fills on the probe, before the setup exists.
• Stop-limit: it caps the price at the cost of sometimes no fill. For an entry that failure is cheap: you have no trade. For a protective stop the same failure leaves you in a losing position (M0.03).

Why a resting limit fills when you are wrong

Touching your price is not a fill. Globex fills each price first come, first served, so every lot ahead of you at 29,500.00 must trade or cancel before yours fills Definition. That is why the hook’s limit missed: the low was a tick above it, and even a touch of 29,500.00 might not have reached your place in line.

Now ask when the whole line at 29,500.00 does get eaten. Mostly when sellers keep coming, and often as one aggressive order that trades through several prices at once, a sweep (M3.11). Your fill then arrives at the moment selling is strongest. The standard market-microstructure model (Glosten and Milgrom, 1985) says a resting order trades against better-informed or more urgent traders often enough that, on average, the price keeps moving against it after the fill: a negative markout PlausibleOther market. On NQ fills this is not yet measured; the box below says what is.

The queue also carries a weak clue about the next tick. When the best bid holds much more size than the best offer, the next tick is more often up: measured on stocks MeasuredOther market, and on NQ in one practitioner’s public study, 65.6% correct one tick ahead PlausibleOne trader's study (pre-registered, but one practitioner, intervals not clustered by session, capped at Plausible). The expected gain is about a third of a tick, less than the spread a market order pays: a timing aid for a limit, not a reason to enter (M0.03).

Scaling, first touch and the entry window

Scaling in means splitting the position across two prices, for example half on the retest (location) and half on the breakout (confirmation). It is legitimate only when the plan fixes both prices, both sizes, the stop for each half and the condition that cancels the second half House rule. Adding mid-trade because “it is working” is averaging up on emotion. Note that the breakout half usually sits farther from the stop, so it carries more risk per contract.

First touch or second touch. The first touch of a level gives a better price; the second gives you one more observation and a worse price. Which one wins on NQ is not yet measured. The rule is procedural: pick one per setup in the plan and log which you took House rule. Your journal (M7.03) will tell you which one your hands execute better.

The entry window. Every entry has a time limit House rule. A pullback entry is valid for the first orderly return; if price chops around the level for an hour and then touches it, that is a new auction you have not analysed. Stale orders get cancelled, not given another chance.

What the data says

Worked example: a retest long (synthetic day)

The plan, written before the open: after a breakout above 29,490, buy the first pullback to 29,500.00. Invalidation below 29,488.00, stop 29,487.75. Target 29,530.00. Window: the first return within 30 minutes of the breakout.

• Plan A, one order. Buy 2 MNQ, limit 29,500.00. Risk: 12.25 points × $2 × 2 = $49.00 before costs. A missed fill means the market did not offer the planned price; that is acceptable.
• Plan B, scaled. Buy 1 MNQ limit 29,500.00, plus a buy-stop for 1 MNQ at 29,516.00 in case price confirms without pulling back. Risk on the limit half: $24.50. Risk on the stop half, with the same stop: 28.25 points × $2 = $56.50, more than double. Cancellation rule: if the limit fills, cancel the buy-stop; if the window closes without the limit filling, the stop half is the whole position.

What happens: NQ drifts back down, trades 29,499.50, and the line at 29,500.00 clears. The limit fills. The trader cancels the buy-stop as planned, and when NQ trades 29,520.00 does not “add the other half at market”. The plan held; the window closed.

The mirror and the failure

Short side. Mirror every order: a sell limit at the planned retest of resistance, say 29,560.00, for the pullback short; a sell-stop below the trigger, say 29,540.00, for the breakdown. A buy sweep up through your sell limit is the same adverse selection, upside down.

The failure. Next session, same plan and level. At 10:41 ET (22:41 ICT) one sell order sweeps from 29,506.00 to 29,496.50 in under a second. Your limit fills in the middle of it. Within a minute NQ trades 29,489.00, and the stop fills at 29,487.75: −$49.00 plus costs. The fill came easily because the selling was strong; one day is an illustration, not evidence PlausibleOther market. Some traders cancel a resting limit when price arrives as a fast sweep and re-check the level; whether that helps is not yet measured, so test it as a house rule House rule.

Prop-firm note

Topstep’s published list of prohibited trading strategies (help-center article, updated 10 June 2026, read 9 November 2026) bans orders outside the best bid/offer and SIM-fill exploitation: strategies built to harvest simulated fills, such as abusing queue position, catching stray fills in a gapped market, or tight brackets and auto-breakeven tuned to the simulator Firm rule. The stated penalty is removal from the program and possible rejection of profit claims; the dated rule card lives in PROP1.04. Evaluation fills are simulated, so a limit fill rate measured there is not your live fill rate; the difference is not yet measured (MKT2.03). Not affiliated with Topstep; rules change, so check the help center before you trade.

Do it on the platform

Our desk sends orders from TopstepX and reads order flow in ATAS X on a second screen.

Common mistakes

Common mistake: using market orders for everything “to make sure I get in”. On a pullback entry you pay at least the spread, plus slippage, for a plan whose whole point was the price. Match the order to the family, and measure what each one costs you in ticks (M10.02).

Common mistake: a limit at the extreme for a rejection entry. It fills on the probe, before the close back inside that defines the setup. The order must serve the trigger.

Common mistake: reading the queue-imbalance result as an entry signal. A 65.6% next-tick hit rate is worth about a third of a tick, less than the spread. That is a timing aid for a limit, not an edge.

Drill: what your fills are telling you

Your journal for two weeks of retest and breakout entries on MNQ:

1. Four breakout entries with buy-stops: triggers 29,516.00, 29,610.00, 29,455.00, 29,702.00; fills 29,516.75, 29,610.25, 29,456.50, 29,702.00.
2. Five retest entries with limits: three filled (two of those hit the stop), two never filled (both would have reached target).

Compute the average stop-entry slippage in ticks and in dollars per MNQ contract. Then say what the limit record does and does not show.

What we don’t know yet

• The markout of a resting NQ limit order filled inside a sweep, versus one filled slowly: not yet measured. Our lab’s sweep study measures the next minute after aggressive trades, not resting fills.
• Whether first-touch or second-touch entries do better on NQ: not yet measured.
• How simulated fills in a prop-firm evaluation differ from live queue fills: not yet measured.

Recap

• A limit promises your price and not a fill; a stop makes a fill very likely and not the price; a stop-limit caps the price and can leave you with no fill.
• Limits for pullback and retest entries, stops for breakout and continuation entries, and a rejection order at the trigger, never the extreme: house pairings, not measured edges.
• A resting limit fills when the line ahead is eaten, often inside a sweep, and the price then tends to keep moving against it: Plausible from microstructure theory, not yet measured on NQ.
• Scaling in, first or second touch and the entry window are written into the plan before the session, with a cancellation rule.
• Order tactics built to exploit simulated fills are prohibited conduct at Topstep.

Further reading

• CME Globex Matching Algorithms (CME documentation): how your entry order is placed in line.
• CME Matching Algorithms Explained (Databento): queue position, and why changing an order’s price sends it to the back.
• Glosten and Milgrom (1985): the model behind adverse selection of resting orders.
• Gould and Bonart, queue imbalance as a one-tick-ahead predictor: the stock-market result.
• Order Flow, Measured (public practitioner study): the NQ queue-imbalance test (claim MR-06).
• Our lab: NQ sweeps continue slightly less often over the next minute than single-price bursts (forward test pending, smaller than a round trip); markouts of resting NQ fills are not yet measured.

Next: M6.07, choosing which entry family is yours, now that you know what each family’s order costs you.

Educational content only — not financial advice. Futures trading involves substantial risk of loss.