Weekly breakout from a sideways box: eight criteria and a stop in the middle

Idea in brief

The author of the Financial Wisdom channel trades stocks on weekly charts only and calls this their main method. The author looks for a stock in an uptrend that has spent at least six weeks in a range with clear boundaries, and calls such a range a box. The signal comes from a week that closes above the upper boundary with a strong gain, near the week's high and on higher volume.

The buy happens at the open of the following week. The initial stop is inside the box, in its middle third. After that the stop is raised: after a downward MACD cross, the low of that week is marked, and the position is closed when it breaks.

There are eight criteria, and most of them have numbers. The video itself is about trading this system through options. The card covers only the stock signal, the options layer is left out.

Why it might work

The author's logic is as follows. A range after a rise shows that sellers could not turn the stock around. A breakout with a close above the boundary, a weekly gain of 5–20% and volume 30% above the previous week reads to the author as participation of large buyers. A close near the week's high means buying lasted the whole week, not just the first days. A new high of at least 10 weeks filters out breakouts inside a wider range.

The author picks the weekly timeframe for less noise and for lifestyle reasons: signals are checked on weekends, there are few trades and costs are small. A stop in the middle of the box gives price room for normal fluctuations and knocks the trade out if the breakout fails deep back into the base.

This video has no verified statistics for the rules. There are two illustrative examples and overall group figures from another video by the author, which does not say which trades are included.

Rules

The box (Finetiq formalization)

// weekly bars, the signal is calculated at the week's close
// author: a sideways consolidation of 6+ weeks with clear resistance and support levels
// Finetiq: box = L weeks before the breakout week, L from 6 to 26
BoxHigh(L)   = Highest(High, L)[1]
BoxLow(L)    = Lowest(Low, L)[1]
BoxHeight(L) = BoxHigh(L) - BoxLow(L)

// Finetiq: "sideways" = box height no more than MaxBoxPct of its top
IsBox(L) = BoxHeight(L) / BoxHigh(L) <= MaxBoxPct         // Finetiq: MaxBoxPct = 15%

// Finetiq: take the longest L from 6..26 for which IsBox(L) is true
Resistance = BoxHigh(L)
Support    = BoxLow(L)

Conditions on the breakout week

C1 = Close > SMA(Close, 20)                           // author: above the 20-week average; Finetiq: simple
C2 = box found, L >= 6                                // author
C3 = MACDLine > SignalLine                            // author: MACD 12 26 9 in the video frames
C4 = Close >= Resistance * 1.01                       // author: close at least 1% above resistance
     AND Close / Close[1] - 1 >= 0.05
     AND Close / Close[1] - 1 <= 0.20                 // author: gain versus the previous week 5–20%
C5 = (High - Close) <= 0.25 * (High - Low)            // author: close near the high, small wick; Finetiq: share 25%
C6 = High > Highest(High, 10)[1]                      // author: at least a 10-week high
C7 = (Close - InitialStop) / Close <= 0.10            // author: "a reasonable stop in percent"; Finetiq: 10%
C8 = Volume >= 1.3 * Volume[1]                        // author: volume 30% above the previous week

Setup = C1 AND C2 AND C3 AND C4 AND C5 AND C6 AND C7 AND C8

Entry

IF Setup
    BUY AT NEXT BAR OPEN                  // author: at the open of the following week, "slightly above the breakout week's close"

// Finetiq: a second reading for testing
//   BUY STOP at Close * 1.005 for the whole following week, CANCEL if not filled

Stop and trade management

// author: the box is split into three equal parts, the stop goes in the middle one
InitialStop = Support + 0.5 * BoxHeight   // Finetiq: middle of the box; on the author's diagram the stop line is inside the middle third
// variants for testing: Support + BoxHeight / 3 and Support + 2 * BoxHeight / 3

Stop = InitialStop
SELL STOP at Stop                         // filled during the week

// author: raising the stop
IF MACDLine crosses below SignalLine      // at the week's close
    Stop = max(Stop, Low)                 // author: low of the cross week; Finetiq: the stop only moves up
// exit when price moves below that low (author)
// the author has no profit target or time exit

Position size (author)

RiskMoney = Equity * 1%..2%               // author: no more than 1–2% of the account per trade
Shares    = RiskMoney / (EntryPrice - InitialStop)
// the author says this about options; for stocks the formula is the same, without the options multiplier

Stock selection (a separate rule)

// author: the main screening is done by the author's box scanner; for a manual search:
Momentum = gain of about 30% over 1, 3 and 6 months   // author
// Finetiq: we read this as Close / Close[4] >= 1.3 OR Close / Close[13] >= 1.3 OR Close / Close[26] >= 1.3
// author: market cap above $25 billion is needed for liquid options
// Finetiq: for trading the stock itself, a turnover liquidity filter instead of market cap

Parameters

What to test

1. Box definition. This is our main formalization. Compare a maximum height of 10%, 15% and 20%, and a length of 6–13 versus 6–26 weeks. Look at 20–30 detected boxes by eye and compare them with the author's examples.
2. Contribution of each criterion. Remove the eight conditions one at a time and compare the average trade in risk units and the number of trades. Volume and the close near the high may not improve trades but only shrink the sample.
3. The 20% upper limit on the gain. Compare breakouts with a weekly gain of 5–20% and above 20%. The author filters out the strongest weeks. Check whether those are the best trades.
4. Stop level. The lower boundary, the middle and the upper boundary of the middle third. Look at the average loss and at the share of stopped-out trades that later went above the entry price anyway.
5. MACD exit. Compare the author's stop raising with a trailing stop on the 10-week average and with an exit after a fixed number of weeks. In systems with rare large profits, the exit often matters more than the entry.
6. The 3.5R and 50% claim. On your own sample, calculate the average profit in units of initial risk and the win rate. If about half the trades are winners and the average profit is about 1R, an edge of the claimed size is not confirmed.
7. Survivorship and market regime. Run it on a universe that includes delisted stocks and stocks removed from the index, and separately in years when the index is below its 20-week average. On the current index membership, breakouts look better than they really were.

Platform notes

TradingView (Pine Script)

• A strategy trades only the chart symbol. Screening boxes across hundreds of stocks is practically unavailable in Pine: the search is done with a screener or by hand, and the backtest runs on a single stock. An overall result across the universe cannot be calculated on TradingView.
• Signal at the week's close, entry at the next week's open: this is the default behavior (process_orders_on_close = false).
• A stop strategy.exit(..., stop = stopLevel) stays active until filled. On a Monday gap down, the fill will be at the open, below the stop. To raise the stop, pass the new level on every bar.
• High excluding the current week: ta.highest(high, 10)[1]. MACD: ta.macd(close, 12, 26, 9).

MultiCharts and TradeStation (EasyLanguage)

• Running across a list of stocks: Portfolio Maestro in TradeStation, Portfolio Trader in MultiCharts. A single chart only tests the signal.
• Highest(High, L)[1] and Lowest(Low, L)[1]. The box length L has to be looped over on every bar.
• The Volume versus Ticks confusion in TradeStation concerns intraday bars. Still, check the weekly volume against exchange data before applying the +30% threshold.
• Sell next bar at Stop stop lives for one bar, that is one week, and is sent again every week. Set the stop as a price: SetStopLoss works in money.
• Second reading of the entry: Buy next bar at Close * 1.005 stop.

MetaTrader 5 (MQL5)

• Stocks in MT5 are most often CFDs with tick volume (iTickVolume). A +30% volume condition on tick volume is already a different rule. Real volume (iRealVolume) is available only for exchange-traded instruments at a few brokers.
• The universe is limited to the broker's list. Scanning hundreds of US stocks in an EA is feasible, but the broker's history has no delisted stocks.
• In iMACD, the signal line is a simple average, not an EMA. Calculate the 9 EMA of the main line (buffer 0) yourself.
• Positions are held for weeks and months. CFDs on stocks charge a swap fee for every night, and it has to be included in the test.

Where the idea can break

• The author finds boxes "with clear levels" by eye or with their own scanner. A mechanical version will find different boxes, and the result may differ noticeably from the author's trading.
• The Intel and Micron examples were chosen for the demonstration. The 3.5R and 50% figures from the other video refer to the author's group and are given without a period, trade count or list.
• The video sells a scanner and a community. There is no publicly available trade report.
• The stop is filled during the week. On a gap down after earnings, the loss will be larger than calculated, and the 1–2% risk per trade will be exceeded.
• The percentage thresholds are tuned for US stocks. On futures and indices a weekly gain of 5–20% is rare, so there will be almost no signals.
• A backtest on the current index membership inflates the result through survivorship. A fair test needs a universe with delistings.
• The options layer from the video (deep ITM calls; according to the author, an 8–10% stop in the stock means a 40–50% drop in the option) is not included in the card. Returns from the options examples do not carry over to the stock rules.