Gap down and a reversal at the 21 EMA or 50 SMA: buying above the bar's high

Idea in brief

Jack Corsellis swing trades leading stocks and enters on three types of candles. One of them is a reversal bar after a gap down. The stock opens below yesterday's close, the open is at or very near the day's low, then price rises all day and closes near the high. Buyers showed up in the first minute and absorbed all the selling.

The candle by itself, without a location, means little. The author looks for such a bar when a stock in an uptrend or in a base pulls back to the 21-day EMA or the 50-day SMA, preferably into an area of prior support and on drying volume. Entry is a stop order above the bar's high, with the stop below the low. The author compares the bar's risk with the 20-day ADR%, the average daily range in percent.

Why it might work

The author's explanation comes from Wyckoff. A gap down at support triggers stops and scares weak holders. If large buyers immediately take that volume and price closes near the high, demand turned out to be stronger than supply exactly where it was expected. The open at the low matters: buying started from the open, not after an intraday sell-off. That is why the author also looks at the hourly chart, where the reversal should start within the first hour.

The author takes the statistics from Thomas Bulkowski. Belt hold bullish, the author's favorite variant, reverses in 71% of cases over 14,542 observations, but by performance the pattern sits in the middle of the list of 103. In Bulkowski's book, the reversal rate describes the direction in which price leaves the pattern. It is not the win rate of trades with this stop and this location, so the rules below do not test these numbers.

Rules

Reversal bar after a gap

GapDown     = Open < Close[1]                           // author: open below yesterday's close
OpenAtLow   = (Open - Low)   <= 0.10 * (High - Low)     // author: open at or near the low; 10% Finetiq
CloseAtHigh = (High - Close) <= 0.25 * (High - Low)     // author: close near the high; 25% Finetiq
Bullish     = Close > Open

GDR = GapDown AND OpenAtLow AND CloseAtHigh AND Bullish
// the five variants from the slide (piercing, belt hold, meeting lines, thrusting, engulfing)
// differ in where the close sits relative to the body of yesterday's candle.
// Finetiq: we take what they all share. The author did not name a minimum gap size

Location and context

EMA21  = EMA(Close, 21)
SMA50  = SMA(Close, 50)
ADRpct = 100 * SMA((High - Low) / Low, 20)     // author: daily range without gaps over 20 sessions; formula Finetiq
Tol    = 0.5 * ADRpct / 100 * Close            // Finetiq: tolerance for touching the average

AtMA = (Low <= EMA21 + Tol AND Close >= EMA21)
    OR (Low <= SMA50 + Tol AND Close >= SMA50)  // author: 21 EMA or 50 SMA, does not use the 10 EMA

Uptrend = EMA21 > SMA50 AND SMA50 > SMA50[20]  // Finetiq: "a leader in an uptrend or a base"

// optional, for testing
DryVolume = Average(Volume, 5)[1] < Average(Volume, 50)[1]   // author: volume dries up on the pullback; numbers Finetiq

Entry and stop

Setup   = GDR AND AtMA AND Uptrend
RiskPct = 100 * (High - Low) / High            // entry at the high, stop at the low

IF Setup AND RiskPct <= ADRpct AND RiskPct <= 5     // author: risk below ADR%, ideally no more than 4-5%
    BUY STOP at High + 1 tick                        // author: entry through the bar's high
    // Finetiq: the order lives for the next bar only, then CANCEL

StopLoss = Low - 1 tick                              // author: stop below the low
EXIT STOP at StopLoss
// the author sometimes accepts risk up to 1.5 × ADR% if the author likes the setup. This is a variant for testing

Trade management (Finetiq)

// the author trails the stop and makes the trade "free", but does not reveal the rules in this video
R = EntryPrice - InitialStop
IF High >= EntryPrice + 2 * R THEN StopLoss = max(StopLoss, EntryPrice)           // breakeven after +2R
IF BarsSinceEntry >= 3 AND Close < EMA(Close, 10) THEN EXIT AT NEXT BAR OPEN      // trailing exit on EMA 10

First-hour filter (author, thresholds Finetiq)

// the author looks at the hourly chart: the reversal should start right after the open
FirstHourLow = low of the 09:30–10:30 ET hour
IF Low < FirstHourLow THEN Setup = false       // Finetiq: the day's low was made after the first hour

Stock selection (author, outside the signal)

The author picks leaders: strong relative strength, an RS line at 52-week highs, a base with tight candles, signs of accumulation when reading the chart from left to right. This is universe selection, and it is largely discretionary. The card describes a signal on a single symbol. The leader scan is done separately, and not on every platform (see the platform notes).

Parameters

What to test

1. Is the average needed. The same reversal bar at EMA 21 or SMA 50 versus the bar without a location condition, and versus the bar at EMA 10. The author rejects EMA 10, but in another video shows an example at the 10 and 21 EMA. If location adds nothing, the setup comes down to a candlestick pattern.
2. Candle tolerances. Open in the bottom 5, 10, 20% of the range, close in the top 15, 25, 35%. Separately, a gap below yesterday's close versus a gap below yesterday's low. If the result holds only on one set, it is overfitting.
3. Candle variants one by one. Belt hold versus engulfing and piercing. Compare the ranking of results with the reversal rates on the author's slide: 71% for belt hold, 63% for engulfing, 64% for piercing.
4. Risk filter. Risk up to ADR% versus up to 1.5 × ADR% and no filter. Look at expectancy in R and at how many setups the filter removes.
5. Gap above the entry level. Count how many entries filled at an open above the bar's high and how much the actual risk grew. For leading stocks, a gap up after a reversal is not rare.
6. Exit. Trailing on EMA 10 versus EMA 21, an exit after 10 days and a 2R target. The author manages trades manually, and a mechanical exit can change the result a lot.
7. First-hour and volume filters. With and without them, separately in rising and falling markets. Leaders in a bear market behave differently from the examples in the video.

Platform notes

TradingView (Pine Script)

• Daily bars for stocks are built on the regular session. On an intraday chart with extended hours, the open and the gap will be different. Calculate the gap on the daily series.
• strategy.entry("L", strategy.long, stop = high + syminfo.mintick) stays active until canceled. The rule limits it to one day, so strategy.cancel is needed on the next bar.
• If the next day opens above the level, the stop order fills at the open. The actual risk is then larger than calculated, while the position size was based on the old level.
• The first-hour filter on a daily chart requires intraday hourly data: request.security_lower_tf with the "60" timeframe.

MultiCharts and TradeStation (EasyLanguage)

• The daily bar depends on the symbol's session template. If the session includes premarket, the open will not match the exchange open, and the gap may disappear. Use the regular session.
• Buy next bar at High + MinMove / PriceScale stop lives exactly one bar. This matches the rule without a separate cancel.
• Set the stop as a price on every bar: Sell next bar at StopLoss stop. SetStopLoss is set in money.
• The hourly series is added as a second data stream. Time is the bar's closing time: the 09:30–10:30 ET hourly bar has Time = 1030.

MetaTrader 5 (MQL5)

• Stock CFDs have only tick volume. The drying volume filter on it is approximate.
• The daily bar is built from the broker's quotes. The open is the broker's first quote after the exchange opens, not the official opening price, so the "open at the low" condition may not match exchange data.
• MT5 charts are built on Bid, while BuyStop triggers on Ask. The entry happens earlier by the size of the spread, while the stop below the low of the Bid bar stays where it is.
• Place BuyStop with an expiration on the next day. A leader scan in an EA is possible if the broker offers the full list of stocks.

Where the idea can break

• The author does not show a backtest of the setup. Bulkowski's statistics describe the candlestick pattern in general, without the location at an average and without this entry and stop.
• The examples in the video are hand-picked: six historical charts and an open trade in GEV. There is one failed example. The author's selection of leaders and the author's reading of the "chart story" are discretionary, and a mechanical trend filter is weaker than the author's eye.
• The author did not cover trade management in this video. The breakeven and the trailing stop here are ours, and the result of swing trades depends heavily on the exit.
• A stop below the low of a long reversal bar can be wide. The ADR% filter removes such cases, and on quiet stocks few setups remain.
• The setup is built on leading stocks in a rising market. In a prolonged decline, a gap down at the 50 SMA more often continues the move.
• Today's list of leaders contains survivorship bias. A test on current leaders will show a better result than reality, so you need historical lists.