BOS and CHoCH: Continuation Breaks, Protected Swings and Confirmation

Video: Break of Structure and Change of Character Explained (BOS & CHoCH)
Creator: Mind Math Money
Original Video: Watch the original video
Difficulty: Intermediate

The Big Idea

Market structure describes the sequence of meaningful highs and lows. In this video’s framework, a break of structure (BOS) extends an established direction. A change of character (CHoCH) breaks the opposing swing that supported the latest continuation move.

These events help organize a chart, but neither guarantees the next move. The introductory claim that a change of character means reversal is stronger than the later explanation of a potential reversal. Treat it as a warning that the previous structure has been damaged.

If You Remember Only One Thing

Identify the swing supporting the latest continuation break before price tests it. A change of character is meaningful only relative to that previously defined structure.

60-Second Summary

• Higher highs and higher lows describe an uptrend; lower highs and lower lows describe a downtrend.
• A bullish BOS exceeds an earlier high in the established bullish structure.
• A bearish BOS extends below an earlier low in bearish structure.
• A bearish CHoCH breaks the low supporting the latest bullish continuation.
• A bullish CHoCH is the corresponding break above a protected high in bearish structure.
• Swings, wick versus close confirmation and time frame must be defined consistently.
• A chart label is not a complete entry-and-exit strategy.

Read the Sequence Before Naming the Break

Imagine highs advancing from 105 to 110 to 115, with intervening lows at 100, 103 and 108. That sequence supports a bullish interpretation. The move above 110 is a continuation break; the low at 108 that leads into the next advance may become the relevant protected low for that later move.

A decline through 108 damages that local sequence. It does not, by itself, establish a long-lasting downtrend. The market could form a range, recover the level or develop a new lower-high/lower-low sequence.

The transcript accidentally calls lower lows and lower highs an uptrend in one sentence. Its surrounding explanation clearly describes a downtrend; this guide uses that intended meaning.

BOS and CHoCH Answer Different Questions

Event	Question	Interpretation
Bullish BOS	Did price extend beyond the relevant high while bullish structure persisted?	Continuation evidence, subject to failure.
Bearish BOS	Did price extend below the relevant low while bearish structure persisted?	Continuation evidence, subject to failure.
Bearish CHoCH	Did price break the low supporting the latest bullish BOS?	Warning that bullish structure has changed.
Bullish CHoCH	Did price break the high supporting the latest bearish BOS?	Warning that bearish structure has changed.

The bearish-to-bullish counterpart is a symmetric explanation added here. The supplied video spends most of its detailed demonstration on bullish continuation and a possible break beneath its supporting low.

Not Every Minor Dip Is the Protected Low

The creator’s definition points to the low of the move that produced the continuation break. That is more specific than whichever small low happens to be nearest the current candle. An internal oscillation can break without overturning the broader sequence.

This distinction is also the main coding difficulty. Two reviewers may choose different swings unless the scale is fixed. A three-bar pivot detector, a rolling range and a discretionary major swing can produce different labels on the same chart.

Support and Resistance Are Areas

The video recommends viewing repeated reactions as zones. A bullish structure can contain support around a cluster of lows rather than a perfectly repeated price. However, a research detector needs a threshold: a wick beneath the zone, a close beneath it, or a close beyond a buffer are different definitions.

The experiment below uses exact pivot prices and completed closes. This is a deliberate simplification of zone-based chart reading. It can classify a small close beyond a level that a discretionary reviewer would still consider inside a broader area.

A Useful Experiment: Continuation or Structural Change?

View Pine Script Backtest Code

//@version=6 strategy("CandleChasers BOS / CHoCH Research", overlay=true, pyramiding=0, initial_capital=100000, commission_type=strategy.commission.percent, commission_value=0.01, slippage=1, calc_on_order_fills=true, process_orders_on_close=false, use_bar_magnifier=true) width = input.int(3, "Pivot bars on each side", minval=1) mode = input.string("BOS", "Event to trade", options=["BOS", "CHoCH"]) rr = input.float(2, "Target R", minval=0.1) hold = input.int(20, "Maximum holding bars", minval=1) buffer = input.int(1, "Stop buffer ticks", minval=1) units = input.float(1, "Fixed shares/contracts", minval=0.001) ph = ta.pivothigh(high, width, width) pl = ta.pivotlow(low, width, width) var float swingHigh = na var float swingLow = na var int highId = na var int lowId = na var int usedHigh = na var int usedLow = na var int bias = 0 var float protectedLow = na var float protectedHigh = na var float fixedStop = na var float fixedTarget = na var int entryBar = na var int processedBar = na lastExit = strategy.closedtrades > 0 ? strategy.closedtrades.exit_bar_index(strategy.closedtrades - 1) : -1 // Fill recalculations must not use the entry/exit bar's final OHLC to discover events. currentEntry = strategy.opentrades > 0 ? strategy.opentrades.entry_bar_index(0) : -1 canProcess = barstate.isconfirmed and bar_index != currentEntry and bar_index != lastExit and (na(processedBar) or processedBar != bar_index) if canProcess processedBar := bar_index if not na(ph) swingHigh := ph highId := bar_index - width if not na(pl) swingLow := pl lowId := bar_index - width up = not na(swingHigh) and close > swingHigh and (na(usedHigh) or highId != usedHigh) down = not na(swingLow) and close < swingLow and (na(usedLow) or lowId != usedLow) changeDown = bias == 1 and not na(protectedLow) and close < protectedLow changeUp = bias == -1 and not na(protectedHigh) and close > protectedHigh int direction = 0 bool changed = false float eventStop = na // Protected-level changes take priority over ordinary swing breaks. if changeDown direction := -1 changed := true eventStop := swingHigh bias := -1 protectedHigh := swingHigh protectedLow := na usedLow := lowId else if changeUp direction := 1 changed := true eventStop := swingLow bias := 1 protectedLow := swingLow protectedHigh := na usedHigh := highId else if up and bias >= 0 and not na(swingLow) direction := bias == 1 ? 1 : 0 bias := 1 protectedLow := swingLow usedHigh := highId eventStop := swingLow else if down and bias <= 0 and not na(swingHigh) direction := bias == -1 ? -1 : 0 bias := -1 protectedHigh := swingHigh usedLow := lowId eventStop := swingHigh selected = direction != 0 and (mode == "CHoCH" ? changed : not changed) if selected and strategy.position_size == 0 and not na(eventStop) candidate = eventStop - direction * buffer * syminfo.mintick valid = direction == 1 ? candidate < close : candidate > close if valid fixedStop := candidate fixedTarget := na entryBar := na strategy.entry("Event", direction == 1 ? strategy.long : strategy.short, qty=units) strategy.exit("Bracket", "Event", stop=fixedStop) if strategy.position_size != 0 int side = strategy.position_size > 0 ? 1 : -1 if na(entryBar) entryBar := bar_index risk = side * (strategy.position_avg_price - fixedStop) if risk > syminfo.mintick fixedTarget := strategy.position_avg_price + side * rr * risk else strategy.close("Event", comment="Invalid gap risk", immediately=true) if not na(fixedTarget) strategy.exit("Bracket", "Event", stop=fixedStop, limit=fixedTarget) if barstate.isconfirmed and bar_index > entryBar and bar_index - entryBar >= hold strategy.close("Event", comment="Time exit") plot(swingHigh, "Confirmed swing high", color=color.red, style=plot.style_stepline) plot(swingLow, "Confirmed swing low", color=color.green, style=plot.style_stepline) plot(bias == 1 ? protectedLow : na, "Protected low", color=color.blue, style=plot.style_linebr) plot(bias == -1 ? protectedHigh : na, "Protected high", color=color.orange, style=plot.style_linebr)

Compare trades following continuation events with trades following protected-level changes, using identical stops, targets and costs. The question is not which label sounds more powerful; it is how each explicitly defined event behaves.

This code tests one objective interpretation of the source’s concept. Mechanical definitions introduced here are ours and are not necessarily the creator’s exact rules.

Layer	Definition
Source concept	Directional continuation breaks and failure of the swing supporting continuation.
Our definition	Confirmed pivots with three bars on each side; completed closes beyond levels; a directional state initialized by its first qualifying break.
Protected swing proxy	The latest confirmed opposite pivot known at a qualifying continuation break, frozen until the next continuation or change.
Shortcuts	One directional state, one position, no retest entries, exact levels rather than zones, fixed quantity.

What This Code Actually Tests

The first qualifying break initializes bullish or bearish state without trading. Subsequent same-direction breaks of unused confirmed swings qualify as BOS. A close through the frozen protected level takes priority and flips the state as a CHoCH. The dropdown selects which event type to trade.

Each pivot becomes available only after its right-side bars have completed. The script does not enter at the earlier pivot bar. It uses the newest confirmed opposite pivot as a proxy for the originating swing; this may differ from the visually selected low or high in the video.

When flat, an eligible event creates a market entry for a subsequent simulated tick, usually the next bar’s open. The stop is the selected opposite pivot plus a one-tick outward buffer. Protection is submitted with the entry, while a 2R target is calculated after the actual fill. The trade also has a 20-bar time-exit rule.

What This Code Does Not Test

It does not reproduce discretionary swing selection, support zones, higher-time-frame alignment, retest entries, liquidity sweeps or all definitions used by other traders. CHoCH terminology is not universal. The script’s state flips immediately after the protected break; it does not require a complete new trend sequence.

Swings can form without strict alternation, and the latest confirmed opposite pivot is only a practical proxy. Thus, results apply to this detector rather than proving the effectiveness of every BOS or CHoCH framework.

Execution Assumptions That Matter

Stops and targets remain fixed during the position. New events can update chart structure but cannot overwrite its bracket. Signals occurring while invested are not queued. A new entry is suppressed on entry-fill and exit bars, and structure processing on those bars is skipped to avoid using a historical fill recalculation’s final OHLC as though it were already known. This deliberate exclusion can reduce the event sample.

If a gap creates an entry beyond the stop, the script requests an immediate close. This is handling an invalid bracket, not guaranteeing bounded loss. Fixed quantity represents exposure; it does not enforce fixed dollar or percentage risk. Costs are illustrative at 0.01% commission per order and one tick slippage, adjustable in Properties.

Bar Magnifier is enabled, but intrabar data coverage varies. Audit fills where stop and target lie within the same bar. Continuous futures charts also require attention to rollover and adjustment conventions; the ES chart in the video is not a universal execution dataset.

Validation status: This script has not been compiled in TradingView or run against historical data here. No performance results are claimed.

Backtest Lab

Experiment	Change	Keep constant
Continuation versus change	BOS versus CHoCH mode.	Instrument, period, pivot width, costs and exits.
Swing scale	2, 3 and 5 bars on each side.	Event mode, holding limit and reward target.
Exit sensitivity	1.5R, 2R and 3R targets.	Same signals, pivot selection and costs.

Compare trade count, average outcome in R, drawdown and holding duration, separately for long and short trades. Preserve an untouched evaluation period. If performance depends on one isolated pivot width, investigate whether the result reflects parameter selection rather than a stable structural effect.

Try It Yourself

On the video’s one-hour chart scale, hide future candles and mark a sequence of highs and lows. Before advancing, identify the last continuation level and its protected opposite swing. Record which close would invalidate the bullish or bearish interpretation.

Repeat with a different swing width. If labels change, explain which smaller turns disappear rather than declaring one chart objectively correct.

Check Your Understanding

• Does CHoCH guarantee reversal? No. It signals damage to the defined prior structure.
• Why can’t a confirmed pivot be traded on its original bar? Its confirmation requires later bars.
• Does every break below a nearby low qualify? Not under the source’s protected-swing definition.
• Why control exits in the comparison? Otherwise different management may explain the result instead of event type.
• Why is the initial state not traded? The baseline needs an established direction before it can label continuation.

Final Takeaway

BOS asks whether an established direction has extended. CHoCH asks whether the swing supporting that direction has failed. The useful research step is to define swings and confirmation timing consistently, then compare events under matched execution rules. A structural label organizes evidence; it does not guarantee the next trade.

---

Risk Disclosure: This material is for educational and research purposes only and is not investment advice. Trading can result in substantial losses. Backtests are hypothetical and depend on costs, liquidity, gaps, order behavior and intrabar assumptions. Actual execution and market changes can materially alter outcomes.

This article is an independent educational companion to the video and is not affiliated with, sponsored by, endorsed by, or authorized by Mind Math Money.