Part XI · Swing trading · Chapter 11N

How swing trades fail

Common failures and their remedies

1 · Ten failures at a glance

The source numbers its ten failure modes, each with a symptom you can see, a mechanism that explains it and a remedy:

The rest of the chapter takes them in three groups: taking trades the framework would refuse, failing to leave when it says leave, and getting the size or the vehicle wrong. Beneath all ten sits one meta-failure, the belief that discipline appears by itself under stress.

2 · Taking trades the framework would refuse

Setup creep

Symptom, failure mode 1: taking trades that do not meet all of the setup's required conditions, because "it's close enough" or because slow weeks make you impatient. This is setup creep.

Mechanism: the framework is designed for selectivity, with each required condition added because the strategy's edge depends on it. Dropping one turns the setup from a "high-probability conditional edge" into a "weak directional bet".

Remedy: make the check physical and written, and accept that most days offer nothing.

• The setup card must be physically present, printed, taped or copied, before you consider the trade.
• Every required condition must be checked and confirmed in writing.
• If checking reveals that one condition is "not quite there", skip, with no exceptions.
• Accept that 70% to 80% of days have no qualifying setup, which is the framework working, not failing. Chapter 13C, drawn from another part of the same source, puts the share at 60% to 70% of days.

Regime obliviousness

Symptom, failure mode 7: taking a setup in a regime that does not support it, such as a Trend Pullback during a Stage 3 to Stage 4 transition, or pre-earnings momentum when the composite signal is +5.

Mechanism: the technical pattern looks similar regardless of regime, so without an explicit regime check the trader sees only the local geometry.

Remedy: read the regime first, and stop there if it fails.

• The composite signal is layer 1 of the hierarchy (Chapter 11B), so check it before any setup-specific analysis.
• Each archetype has an explicit composite threshold: +30 or more for the Trend Pullback, +20 or more for pre-earnings momentum, and so on (Chapters 11H and 11I).
• If the composite is below the threshold, do not run the rest of the analysis, because the answer is no trade.

The conviction score of Chapter 13B shows how hard the regime bites. Its worked Trend Pullback on SPX scores 0.527, high conviction, in a calm regime with the composite at +50. The same setup scores 0.040 in a stress regime with the composite at −30, where the regime factor falls to 0.20 and execution and sizing weaken with it.

Indicator stacking

Symptom, failure mode 8: "Five indicators confirm this trade!", when the five all measure the same thing. This is multicollinearity in practice.

Mechanism: RSI(2), IBS, Bollinger %B, Williams %R and the stochastic oscillator all measure how stretched the price is relative to its recent range. When the underlying is stretched, their confluence is mathematically guaranteed, as Chapter 11C derives. The worked example below computes all five on one set of bars.

Remedy: count functions, not readings.

• Use one indicator per pedagogical function, as Chapter 11C sets out: one each for momentum, trend, volatility and range.
• If you cannot say what feature of the data each indicator uniquely measures, drop the duplicates.
• Genuine confluence is across different functions, such as momentum, trend, volatility and volume, not across reparameterizations of one.

The source is not consistent about one pair. Its setup chapter calls IBS with RSI(2) genuine confluence, while this failure mode lists the two among five measures of one quantity (Chapters 11C and 11H). Chapter 11C suggests settling it on your own data, by measuring how closely the two move together.

Trading without a written setup card

Symptom, failure mode 10: "I see a setup forming", and a trade taken without explicit reference to the setup's definition.

Mechanism: the criteria you match against are the criteria you can remember. Without an external reference, those become the criteria that conveniently match the current chart.

Remedy: read the card, then write the entry down before sending it.

• Setup cards (Chapter 13F) must be physically referenced before every trade.
• Write a confirmation: "Conditions A, B, C, D met: entering at $X with stop at $Y, time stop at day Z".
• The act of writing slows the decision by 20 to 30 seconds, which is exactly long enough to catch impulse trades.

3 · Failing to leave when the rules say leave

Holding through the time stop

Symptom, failure mode 2: the setup's time stop arrives without the profit target being hit. With the position at a small profit or a small loss, you hold "to see if it works".

Mechanism: time stops exist because, in the source's account, empirical work shows that swing setups which do not pay within their typical horizon have low forward expected value. The probability distribution of "this is a delayed winner" against "this is a slow loser" is heavily weighted toward the latter.

Bayes' rule shows why the weighting shifts: a trade still open at the time stop has already told you something about itself.

Remedy: make the time stop mechanical, and treat a new thesis as a new trade.

• Time stops are mechanical: close at the specified day's close, regardless of P&L.
• If you genuinely have a new thesis, such as "this is becoming a multi-week trade, not a 5-day swing", close the original position and re-enter under the new framework. Never roll the same trade forward without resetting the analysis.

Drag the share of winners paid in time toward 100% and watch the bar on the right. The green share, the chance that a trade still open is a delayed winner, falls far below the base rate.

win rate

60%

winners paid in time

85%

losers still open

50%

cases example check

delayed winners = 0.60 × 0.15 = 0.090

slow losers = 0.40 × 0.50 = 0.200

P(winner | open) = 31%

More likely a slow loser. A trade open at the time stop is a winner 31% of the time, against a base rate of 60%: the wait itself is evidence against it.

Stop migration

Symptom, failure mode 3: the position is approaching its pre-committed stop, and you move the stop further away "to give it room". This is stop migration.

Mechanism: Loss aversion, the central finding of prospect theory, is at work. At the stop, you are psychologically drawn to avoid the realized loss by deferring it. A form of the disposition effect (Chapter 13A), stop migration converts small losses into large ones systematically.

Chapter 13D's Example 9 prices one migration: you are long 100 SPY at $580 with a stop at $575, and the price is $576.50. Kept, the stop triggers and fills at $574.20, a loss of $580. Moved to $572, it fills at $570.50, a loss of $950, which is 64% more than the plan.

Remedy: stops move one way only, and repeated migration triggers the cool-off.

• Stops, once set, can only be moved toward profit, trailed up on winners as a trailing stop (Chapter 11E), and never away from it.
• If you genuinely have new information, close the existing position at market and re-enter with a new analysis, rather than migrating the stop.
• The 25% drawdown and 30-day cool-off of Chapter 13C is the structural defense when stop migration has happened repeatedly. A drawdown of 25% from peak to trough (PTP) halts all discretionary tactical trading for 30 days.

Skipping the exit before earnings

Symptom, failure mode 5: a pre-earnings momentum trade is going well and earnings approach. "Just hold through; the momentum is strong."

Mechanism: Confirmation bias combined with sunk-cost framing. The position has been profitable, so the trader projects its continuation through the binary event.

Remedy: the exit is the trade, so take it on schedule.

• The exit-before-earnings rule is the defining discipline of the pre-earnings momentum trade (Chapter 11I).
• Close mechanically on T-1 at the latest, the last trading day before the report, regardless of how good the trade looks.
• Single-stock event risk requires defined-risk options, not naked equity, if the position is held through, which makes it a structurally different trade.

The arithmetic is short: a stock with an implied move of ±9% and a stop 3.5% below entry can gap through the stop on a move of ordinary size. A 9% gap against the position costs $9/3.5 = 2.6$R, where the plan allowed 1R.

4 · Getting the size or the vehicle wrong

Sizing inconsistency

Symptom, failure mode 4: a larger position size after winning streaks and a smaller one after losing streaks.

Mechanism: Overconfidence after wins and loss aversion after losses push sizing in the wrong direction. The first adds risk when conditions have not changed, while the second removes it when nothing about the setup's edge has changed. If the trades are independent, a trade after three wins has the same win rate as any other, so extra size buys variance, not edge.

Remedy: size by formula, at the same percentage, whatever the recent results.

• Position sizing is mechanical: $n \le \lfloor B_{\text{exp}}/L_{\max} \rfloor$, where $B_{\text{exp}}$ is the dollar loss budget for the trade and $L_{\max}$ the loss per unit at the stop (Chapter 13C).
• The loss budget $B_{\text{exp}}$ is the same percentage of the portfolio regardless of recent P&L.
• Recent wins do not permit larger sizing, just as recent losses do not require smaller sizing.
• The only legitimate changes to sizing are regime-based reductions, when the composite weakens, and the drawdown rule's kill switches.

Correlated overconcentration

Symptom, failure mode 6: three Mag-7 longs running at once, sized as if independent, or three Compression Breakouts in growth stocks.

Mechanism: the setups are nominally different, but the underlying risk factor, Mag-7 momentum or growth-stock dispersion, is the same: pairwise correlations during stress can hit 0.9 or more. Three positions of equal risk at a correlation of 0.9 have a combined standard deviation of $\sqrt{3 + 6 \times 0.9} = 2.9$ times one position's, against $\sqrt{3} = 1.7$ times if they were independent (Chapter 11L).

Remedy: budget the factor, not the ticket.

• Treat correlated positions as one position for risk-budget purposes.
• The portfolio-level loss budget bounds the total, not the per-trade budget times the number of positions.
• Cap simultaneous correlated positions, for example at two long Mag-7 swing positions at any time.

A horizon mismatch

Symptom, failure mode 9: an expiry that does not fit the thesis. Buying 7 DTE options for a 14-day swing thesis is one form, and buying 90 DTE options or LEAPS for a 3-day pivot trade is the other.

Mechanism: the option's expiry does not match the thesis horizon. Either it expires before the thesis develops, or it carries unnecessary cost for a fast resolution.

Remedy: date the vehicle from the thesis.

• Apply the DTE matching principle of Chapter 11M, which sets each setup's expiry from its holding period.
• Option expiration should be at least the thesis horizon plus a buffer of 2 to 3 trading days.

5 · The meta-failure: assuming discipline appears under stress

The deepest failure mode is implicit: assuming that you behave more carefully in stress than in calm. Most retail traders believe this about themselves, a belief that the source says the data systematically refute. Chapter 13A sets out the behavioral evidence behind the claim.

Discipline exists only when it is binding under stress. It is binding only when it was written down in calm. The setup cards, the kill rules, the time stops and the pre-committed exits exist as external constraints on your behavior, not as suggestions for it. The professional bias is to design the framework so that, even under maximum behavioral stress, it defaults you toward the right action. That is what discipline is: not willpower, but structural design.

6 · Check yourself