Clean Execution

Execution is a process, not a moment

Most traders think of execution as the click. In reality the click is the last 2 seconds of a process that started an hour earlier. If the hour was done properly, the click is boring. If it wasn't, the click is a guess.

Module 02 gave you the setup model and Module 03 gave you size and limits. This module covers the part in between: getting from "this is an A-grade idea" to a correctly placed, managed and reviewed order. Every decision that can be made before the market moves should be made before the market moves.

The pre-market routine

Call the start of your killzone T. If you trade NY AM from 08:30 ET, T = 08:30 ET (16:30 GST summer / 17:30 GST winter). The routine has three checkpoints.

T−60: bias and draw on liquidity (15–20 minutes)

1. Daily and 4H structure: one sentence, as in Module 02 ("Daily bullish, protected low 19,800").
2. Premium or discount of the current dealing range.
3. The draw on liquidity (DOL) for today: the one pool price is most likely to reach.
4. Overnight action: Asia and London highs and lows, and whether either was swept.

T−30: levels and news (10 minutes)

1. Mark PDH/PDL, session highs and lows, and any unfilled 1H/15m FVGs between price and the DOL.
2. Check the economic calendar. Write down every high-impact release time. Apply the news flat rule (see Module 03).
3. Pick at most two levels where you would act. More than two means you have no plan.

T−10: state check and orders staged (5 minutes)

1. State check from Module 03. If you are not at your baseline, size down a grade or sit out.
2. Confirm today's risk: daily stop, max trades, the dollar value of 1R.
3. Load your bracket template (Chapter e3) on the correct contract and account. Do not place it yet.
4. Read your trade scripts out loud once.

Session timing: when to be at the screen

Be at the screen for the window you trade, not for the whole day. Dubai has no daylight saving, so GST times shift by one hour when the US changes clocks: "summer" means US daylight time (EDT), "winter" means US standard time (EST).

From Dubai, the NY AM window lands in the early evening, which is convenient but competes with dinner, calls and social plans. Protect it like a meeting. London falls in the late morning, which suits a trader who wants to finish before lunch.

Pick one primary window and, at most, one secondary. Your T−60 routine should start an hour before your primary window opens. The clock lab below shows today's windows in Dubai or New York time and lights up whichever one is active, with daylight saving handled automatically.

Writing the trade script

A trade script is an implementation intention: a plan in the form "if situation X, then I do Y". Peter Gollwitzer's research (1999) found that if–then plans make people noticeably more likely to follow through on a goal than simply intending to, because the decision is made in advance and triggered by the situation.

The EdgeFlow script has six slots: level, time, trigger, entry, stop, target.

Write the opposite script too: "If price instead runs above 20,150 first, then I do nothing until a new script is written." The second script is just as important. It tells you what not to do.

Two scripts per session is the maximum. Each must fit on two lines. If you can't write it, you can't trade it.

The wait

Most of a session is waiting. Waiting is where boredom trades come from: entries taken not because the script triggered, but because nothing happened for 40 minutes.

What to do while price is away from your levels:

• Watch the 15m and 5m only. Close the 1m until price is within roughly one ATR of your level.
• Update your journal's pre-trade notes: what has happened since T−10.
• Set price alerts at both script levels and step away from the chart if needed.

Rules that stop boredom trades:

1. No script, no order. If the setup isn't written down, it doesn't exist.
2. No new levels after T. Levels added mid-session are almost always justifications for a trade you already want.
3. Killzone ends, you stop. At 11:00 ET (19:00 / 20:00 GST) the NY AM session is over for you, traded or not.
4. A zero-trade day is a completed day. It counts as following the plan in your journal.

The one-page execution flow

Print this and keep it beside the screen. Every trade passes through seven steps, in order.

If any step fails, the flow stops. You don't skip from Bias to Order because price is "moving fast". Fast markets are exactly where skipped steps cost the most.

The three EdgeFlow entries

All three entries happen after the trigger in your script: a liquidity sweep followed by a market structure shift with displacement (Module 02). They differ only in where you get in.

1. FVG entry (default). After the MSS, mark the fair value gap created by the displacement leg. Place a limit order inside that gap. Stop beyond the sweep extreme plus your ATR buffer (see Module 03). Valid until a candle body closes through the far side of the gap.

2. Order block entry. Mark the last opposing candle before the displacement (for a long, the last down-close candle). Enter on a limit at the top of that candle's body, or its mean threshold (50% of the body). Stop beyond the order block's wick or the sweep, whichever is further. Use it when the displacement left no clean FVG, or when the FVG and order block overlap.

3. MSS-retrace entry. Enter on the retrace to the broken swing level itself: the old swing high (for a long) that price just closed above. Old resistance becomes support. Use it when the FVG is tiny or already filled, and the broken level sits close to the sweep so the stop stays tight.

Aggressive vs conservative entries

Within an FVG there are three standard ways in:

• Aggressive (edge): limit at the near edge of the gap, the first price the retrace touches. Highest fill rate, smallest R.
• Consequent encroachment (CE): limit at the 50% midpoint of the gap, the ICT term for its "fair" price. Better R, but price often turns before reaching it.
• Confirmation: wait for a candle (usually 1m or 5m) to close inside the gap and hold, then enter at market. Filters some failures, but your price is worse than CE and you still miss trades that never pulled back far enough to close inside.

The deeper you wait, the better the R on the trades you get, and the fewer trades you get. The missed trades are usually not random: they are often the strongest moves, because strong moves don't retrace far.

The CE entry has the best R per trade and the worst total. That's the trade-off in one line: R per trade is not the same as R per setup. Your backtest, not your preference, decides which entry you use.

Choosing your entry

Use this decision rule until your own data (from the entry lab and your journal) says otherwise:

1. A+ setup with strong displacement: edge entry. Strong moves rarely give you CE. Missing them costs more than the extra R gains.
2. A setup, normal displacement, FVG wider than about 0.5× the 5m ATR: CE entry. The gap is wide enough that the edge entry's R is poor.
3. B setup, or choppy conditions: confirmation entry, or skip. Let price prove itself.
4. Gap narrower than a few ticks: treat the FVG as a level, not a zone, and use the MSS-retrace entry instead.

One option many traders use is a split entry: half size at the edge, half at CE. Your average entry and R land between the two, and you are never fully out of a runner. Total risk must still equal your planned R for the grade, so calculate the stop distance from the average price.

Entry timeframe vs context timeframe

Module 02's multi-timeframe rule still applies: higher timeframe for direction, lower for timing. For execution, map it like this:

Your entry gap must sit inside the context picture. A 1m FVG pointing long while the 15m is displacing down is not an entry. It's a pause in the move against you.

Stops are set on the trigger timeframe (beyond the sweep), not on the entry timeframe. A 1m stop under a 1m candle is too tight for a 5m idea and will be hit by normal noise.

Limit, market or stop-entry

Match the order type to the entry style:

• Limit order: edge, CE, order block and MSS-retrace entries. You decide the price in advance and wait. This is the EdgeFlow default.
• Market order: confirmation entries only, placed at the close of the confirming candle. Accept a few ticks of slippage.
• Stop-entry (buy stop above / sell stop below): rarely. Only for a break-and-go entry where your script says "enter if price closes above X and continues". It guarantees you buy into strength, usually at the worst price of the move.

A limit order also enforces discipline. Once it's placed with a bracket, the trade either happens at your price or not at all.

The same logic applies outside index futures. On EURUSD, an FVG entry is a buy limit inside the gap with the stop beyond the sweep, exactly as on NQ. On BTC perps, limit orders usually also pay lower (maker) fees than market (taker) orders, so the default limit entry saves you money on every trade, not just on price.

Missed entries: the no-chase rule

The no-chase rule: if price leaves your entry zone without filling you, you do not enter at a worse price. No market order "to get in", no moving the limit up.

Chasing breaks the trade's maths. In the example above, chasing to 20,040 means risk 65 points and reward 80, so R drops from 2.22R to 1.23R, and your stop is now further from the real invalidation than your plan allows.

A re-entry is valid only if all four are true:

1. Price forms a new sweep and MSS, or retraces into a new FVG created by the continuation move.
2. The new stop and target still give at least your minimum R.
3. You are still inside your killzone.
4. It counts toward your max 3 trades per day.

Otherwise, the missed trade goes in your journal as "missed, rules followed", which is a good result.

Anatomy of a bracket order

A bracket order is three orders that work as one:

1. Parent (entry): usually a limit order at your entry price.
2. Protective stop: a stop-market order beyond your invalidation.
3. Target: a limit order at your take-profit.

The stop and target are linked as OCO (one-cancels-other): when one fills, the platform cancels the other. Both only go live once the parent fills. Most futures platforms (Tradovate, NinjaTrader, Rithmic-based platforms) call a saved bracket an ATM strategy or bracket template.

Why it matters: from the moment you are filled, the trade is fully defined. If your internet drops, the stop and target are already sitting at the exchange or broker server.

Building the order ticket

Every ticket follows the same order of operations: stop distance first, size second, click last.

1. Entry price from your script.
2. Stop price from your script. Stop distance in points = entry − stop (long) or stop − entry (short).
3. Dollar risk per contract = stop distance × point value.
4. Contracts = planned dollar risk ÷ risk per contract, rounded down.
5. Target price, and check R = reward ÷ risk.

Illustrative prices.

Always round down. Rounding 7.5 up to 8 MNQ turns $300 risk into $320. Small, but it compounds across a month and breaks prop drawdown maths.

Stop-market vs stop-limit for your protective stop

A stop-market order becomes a market order when price touches your stop. You are guaranteed to get out, but not at an exact price.

A stop-limit becomes a limit order at a set price when triggered. You are guaranteed a price no worse than your limit, but not guaranteed a fill. If price gaps or moves through your limit in a fast candle, the order sits unfilled while the trade runs further against you.

EdgeFlow rule: protective stops are always stop-market.

Picture a short GC position with a stop-limit at 2,410.0 / limit 2,410.5 on CPI morning (illustrative). Price jumps from 2,405 to 2,418 in one second. Your stop triggers, but no sellers exist at 2,410.5, so you are not filled. You are now 8 points (8 × $100 = $800 per GC contract) beyond your planned exit and still in the trade. A stop-market would have filled you at, say, 2,411 with one point of slippage.

Slippage, partial fills and fast markets

Slippage is the difference between your order price and your fill. On liquid index futures in normal conditions it is often zero to one tick on a stop. In fast markets (data releases, the 09:30 ET cash open, thin overnight hours) it can be several points.

Partial fills happen with limit orders: you wanted 5 contracts, price touched your level and only 2 filled before it left. Rules:

• Most platforms attach the bracket to the filled quantity automatically. Check that the stop shows 2, not 5.
• Do not "complete" the position with a market order. That is chasing (Chapter e2).
• Manage the 2 contracts exactly as planned. Your R is the same; only the dollar size is smaller.

Budget for slippage before you size. If your journal shows your stops on MNQ slip 1 point on average, size as if the stop were 1 point wider. On the ticket above that means 31 points × $2 = $62 per contract, and $300 ÷ $62 = 4.8, so 4 MNQ rather than 5. It feels conservative, but it makes your real losses match your planned R, which is what prop drawdown limits measure.

Fast markets: the news flat rule from Module 03 is your main protection. Beyond that, avoid entries in the first minute after 09:30 ET, and expect wider spreads on CL and GC around their own scheduled data.

Platform setup that prevents errors

Most costly platform mistakes happen once and are fully preventable with one-time setup.

• Default bracket templates. Save one template per market with stop and target in ticks as placeholders. Every order you send carries a stop by default.
• Hotkeys off, or with confirmation. A stray keypress should never send an order. If you use hotkeys, enable an order confirmation dialog.
• Correct contract month. Index futures roll quarterly (March, June, September, December). Most volume moves to the next contract about a week before expiry. Check your chart and your order ticket show the same contract month. Trading the expiring month after the roll means thinner liquidity and wider fills; charting one month and trading another means every level is off by the spread between the two contracts, which on NQ can be well over 100 points (illustrative).
• Account selector. If you run several accounts (personal plus prop evaluations), check the selected account before every session. Name accounts clearly in the platform if it allows.
• Max position limits. Set a platform-level maximum contracts per order and per position at or below your largest planned size. If your prop account allows 10 MNQ and you trade 5, set the limit to 5.

The pre-click check

Five things, five seconds, every order. Say them silently as you look at the ticket:

1. Account: the right account is selected.
2. Instrument: right market and right contract month (MNQ, not NQ).
3. Direction: buy or sell matches the script.
4. Size: contracts match your calculation for this grade.
5. Bracket: stop and target are attached and on the correct side of the entry.

If any one is wrong, cancel and rebuild. Never fix a live order while price is moving toward it.

After the fill, do one more glance: the position window shows the right quantity, and both the stop and target appear as working orders. If the stop is missing for any reason, place it immediately at your script price before doing anything else.

Decide management before entry

Once you are in a trade, every tick changes how you feel about it. That is the worst possible moment to make decisions. So EdgeFlow traders make every management decision while flat, and write it in the trade script from Chapter e1.

Your management plan answers four questions before the order goes in:

1. Where is TP1? The first opposing liquidity (a 5m swing high for a long, a swing low for a short).
2. What moves the stop to breakeven? TP1 filled, or a new protective swing formed.
3. What do you trail behind? 5-minute swings or newly formed FVGs.
4. Where is the final target? The draw on liquidity you identified in your pre-market routine.

If you can't answer all four, you don't have a trade yet. You have an opinion.

The EdgeFlow management rules

These are the default rules. Use them for at least 50 trades before changing anything, so your journal data means something.

The breakeven rule is the one most traders get wrong. Moving to breakeven at a fixed +0.5R feels safe, but normal retracements often come back to test the entry FVG. A fixed early breakeven turns many of your best trades into scratches. A new swing low above your entry, by contrast, is the market showing you a new invalidation point.

Time stops

Price is not the only thing that can invalidate a trade. Time can too. ICT setups depend on displacement inside a killzone; if neither happens, the reason for the trade has gone.

EdgeFlow time stops:

• No displacement within 6 five-minute candles (30 minutes) of the fill: close the position, or cut it to half if it is above entry. A good setup usually moves quickly.
• Killzone ends (11:00 ET = 19:00 GST summer / 20:00 GST winter): tighten the runner's stop behind the last 1-minute swing.
• Hard flat at 11:30 ET (19:30 GST summer / 20:30 GST winter) for NY AM trades, unless you have a written plan for a PM session trade. For London trades, the equivalent cut-off is 05:00 ET (13:00 GST summer / 14:00 GST winter).

A time-stop exit usually lands somewhere between −0.5R and +0.5R. That is the point: it removes the slow, drifting trades that end as full losses at lunch.

Valid early exits vs fear exits

You will sometimes exit before the stop or target. That is fine when the market gave you a written reason. It is a leak when your stomach gave you the reason.

Valid early exits (close the position, or at least the runner):

• A candle body closes back through the entry FVG on your entry timeframe. The gap that justified the trade has failed.
• Opposing displacement plus an MSS on the 5-minute chart. The other side has taken control.
• High-impact news is approaching inside the news flat window from Module 03 (5 minutes before CPI, NFP at 08:30 ET or FOMC at 14:00 ET).

Fear exits (do not act on these):

• One large red candle that doesn't close through your FVG.
• Giving back some open profit before TP1.
• "It feels heavy" with no structural change.

Shefrin and Statman (1985) named the disposition effect: people sell winners too early and hold losers too long. Odean (1998) found this pattern in thousands of real brokerage accounts. Fear exits are the disposition effect in intraday form. The fix is a short list of valid exits, written down, and nothing else.

Walking through a live trade

Here is the trade from the worked example, candle by candle, on the 5-minute chart. Setup: NQ AM session, London low at 19,970, draw on liquidity is PDH 20,060. Each row shows what happened and the decision the rules give.

Illustrative prices.

Notice how few actions there are: one entry, one partial, one move to breakeven, two trails, one exit. Six decisions across an hour, all pre-written. If price had reversed at 10:30 and hit the 19,995 trail, the trade would have closed at 0.75R + (5 × 15 × $2 = $150 = 0.375R) = +1.125R. Still a well-managed trade.

In the manage lab you replay trades like this one and pick your action at each candle. The lab scores you against the rules, not against the outcome.

Over-management: the most common profit leak

Over-management means touching the trade without a rule telling you to: moving the stop early, taking extra partials, closing on noise. Each action feels prudent. Added up, they are expensive.

The numbers will be different for you. That is why Chapter e6 shows you how to measure it from your own MAE data instead of guessing. But the shape of the result is typical: early protection saves small amounts on losers and costs large amounts on winners.

Why execution leaks matter more than strategy

A tested setup with a small positive expectancy is a thin edge. Execution mistakes don't need to be big to wipe it out. They just need to be frequent.

Ten mistakes in 60 trades is a 1-in-6 error rate. Most traders who keep honest logs find their rate is higher. This is why you fix execution before you look for a better strategy: a new strategy run with the same leaks will leak the same way.

Barber and Odean (2000), in "Trading is hazardous to your wealth", found that the most active retail traders earned the lowest net returns, largely because of costs and poor decisions. Excess, unplanned activity is a leak on its own.

The eight execution leaks

The R costs below are illustrative estimates to show the typical size of each leak per occurrence. They are not research findings. Your own costs come from your journal, using the method in the next section.

Illustrative estimates, not measured statistics.

Chasing is worth one more line of maths, because it hides well. Plan: entry 19,980, stop 19,960, target 20,040. That is 60 ÷ 20 = 3R. Chase in at 19,990 with the same stop and target: risk 30, reward 50, so 50 ÷ 30 = 1.67R. If you keep risk at 1R, every winner pays 1.33R less, and the losers still cost the full 1R.

Measure your leaks

You don't need to guess the costs. For every trade in your journal, record two numbers:

• Plan R: what the trade would have made if you had followed every rule exactly (entry, stop, size, partial, trail, target).
• Actual R: what it really made.

The leak is the difference: leak = plan R − actual R. Tag each leak with which of the eight caused it. Total by tag and you have a price list for your own bad habits.

Illustrative figures.

The rules would have made +4.5R. You made +0.9R. Early profit-taking cost 2.0R, revenge 1.0R, the moved stop 0.6R. You now know which leak to fix first, and it isn't the one that felt worst.

For a missed or skipped trade that would have followed the rules, plan R is the setup's result and actual R is 0. Count it. Hesitation is a leak too.

The rule-break log

The rule-break log is a separate page in your journal with one line per broken rule, written the same day. Keep it short so you actually fill it in:

• Date and time (ET, plus GST if that helps you spot fatigue patterns).
• Leak tag (one of the eight).
• Cost in R (plan R − actual R).
• State before the break (the Module 03 state check score).
• Trigger: one sentence on what you felt or saw just before.

The trigger column is where the value is. After 20 entries, patterns appear: revenge trades after a loss before 10:00 ET, early exits when you're up on the day, chasing on high-volatility opens. Tendler (2021) calls this mapping your mental game: each leak has a recognisable lead-up you can learn to spot early.

Fixing one leak at a time (30-day focus)

Trying to fix all eight leaks at once usually fixes none of them. Pick the single leak with the highest total R cost from your log and work on it for 30 days.

1. Pick one leak with the largest R total over your last 20–40 trades.
2. Write an if–then plan. Gollwitzer (1999) showed that implementation intentions ("If situation X, then I do Y") make follow-through much more likely than general goals. Example: "If my runner is open and no valid exit has printed, then I take my hands off the mouse and read my exit list aloud."
3. Add a hard control where possible: brackets locked, platform loss limits, alarms at window end.
4. Count weekly: number of occurrences and total R cost.
5. After 30 days, if the leak is down by at least half, move to the next most expensive one.

Tendler's inchworm idea fits here: you don't jump from your worst execution to your best. You push the bottom end up a little at a time, so your worst days get less bad. That's where most of the money is.

A sample 30-day card for leak 5 (early profit-taking) looks like this: target "zero runner exits without a valid exit printed", control "TP1 and final target set in the bracket before entry, runner closed only by the bracket or a listed exit", review "every Friday, count cut runners and add up plan R minus actual R". If week one shows four cut runners and week four shows one, the fix is working, even if the month's P&L was flat.

Grade the execution, not the outcome

Any single trade's result is mostly noise. Mark Douglas (2000) put it plainly: a setup is an edge over many trades, and any one trade can go either way. If you judge yourself by wins and losses, you'll learn the wrong lessons.

So EdgeFlow separates process (did you follow the rules?) from outcome (did it make money?):

The "lucky win" box is where bad habits are made. A moved stop that comes back and wins teaches you that moving stops works. Log it as a rule break (Chapter e5) and score it low.

The execution scorecard

Score every trade out of 10, five areas, 0, 1 or 2 points each.

Aim for an average of 8 or above across 20 trades. Below 7 means execution, not strategy, is your bottleneck. Steenbarger (2006) argues that structured self-review like this, done consistently, is what separates traders who improve from those who just accumulate screen time.

The screenshot protocol

Your memory of a trade is biased by how it ended. Screenshots taken in the moment are not. Take four for every trade, and two for every skipped A-grade setup:

1. Before entry (context): 15m or 1H chart with bias, draw on liquidity and key levels marked.
2. At entry (trigger): 5m or 1m chart showing the sweep, MSS and FVG, with the order levels visible.
3. At exit: the same chart showing where you got out and why, in one line.
4. End of session: the full move, so you can see what the trade was really worth.

Annotate every screenshot with the time (ET), your score and a sentence. "Exited runner at 20,030, fear, no valid exit" is more useful than ten minutes of thinking about it afterwards.

MAE and MFE: what your trades tell you about stops and targets

Two numbers for every trade, both measured in R from your entry:

• MAE (maximum adverse excursion): the furthest the trade went against you before it closed. If your stop is 20 points and price went 8 points against you, MAE = 8 ÷ 20 = 0.4R.
• MFE (maximum favourable excursion): the furthest the trade went in your favour. If price reached 60 points in profit, MFE = 60 ÷ 20 = 3.0R.

The idea was popularised by John Sweeney in the 1990s. Read together, these numbers answer the questions you usually argue about from feel:

• Are my stops too wide? If almost no winning trade ever had MAE above 0.5R, you might tighten. Test this carefully across 50+ trades; a stop tighter than structure stops being a structural stop.
• Are my stops too tight? If many losers were stopped at −1R and then went on to hit target, your buffer may be too small.
• Are my targets right? If MFE often runs well past your final target, your draw on liquidity may be too close. If trades rarely reach TP1, your TP1 is too far.
• Is my management leaking? Losing trades with MFE above 1R mean you had a winner and gave it all back.

Analyse your MAE / MFE

Here are six illustrative trades. All values in R.

Illustrative figures.

What this small sample suggests:

• Winners' MAE never went beyond 0.5R. A first sign that the stop has room, but six trades prove nothing. Keep collecting.
• Trade 4 reached 4.0R but banked only 1.0R. Check the screenshots: was that an early exit or a planned target that was too close?
• Trade 5 reached +1.2R and then lost the full 1R. TP1 was at 1.5R so it never paid. If this keeps happening, TP1 may be sitting beyond the first real liquidity.

In the mfe lab you enter your own trades and see them as two scatter plots: MAE of winners vs losers (stop fit) and MFE vs result (target fit and management).

Execution efficiency

Execution efficiency tells you how much of the available move you actually kept. In plain words: of all the R that your winning trades offered, what share did you bank?

Efficiency = captured R ÷ available R, where captured R is the total result of your winners and available R is the total MFE of those same winners.

No one captures 100%, because you can't sell the exact top. As a rough EdgeFlow guide: below about 40% suggests early exits or targets set too close; 50–70% is healthy for a partial-plus-runner approach. What to change depends on the cause. If the screenshots show fear exits, fix the leak (Chapter e5). If they show rules followed but targets too close, review where you set the draw on liquidity.

The 20-trade execution review

Every 20 trades, block out 60 minutes and run this review. Twenty is enough to see patterns but short enough to catch a leak before it costs a month.

1. Sort into the 2×2 grid. Count each box. More than 3 lucky wins is a warning.
2. Average the scorecard. Find the area with the lowest average score.
3. Total the leaks from your rule-break log by tag, in R.
4. Plot MAE and MFE in the mfe lab. Note one observation on stops and one on targets.
5. Calculate efficiency for the 20 trades.
6. Write one change for the next 20 trades: one leak to fix or one rule to test. Only one.
7. Compare with the previous review. Did the change work, measured in R?

Van Tharp's point about R-multiples applies here: once every trade is expressed in R, you can compare across markets, sizes and months. That turns your review from storytelling into data.

Review the moduleGlossary flashcards →