Delayed reaction after a gap: red-to-green and consolidation

Idea in brief

Both versions in this card answer one question: how to avoid buying on the gap day itself. On a gap, the stop to the day's low is often huge, and after a strong first day the stock often pulls back.

Pradeep Bonde looks for an event with a strong catalyst after which the stock failed to hold the gain on the first day. HIMS opened and closed lower, Netflix declined all day after a gap up, Reddit gave back 41%. Bonde writes down the catalyst and keeps the stock on a list for at least a month. On one of the following days the stock opens below yesterday's close. When price moves above it during the day (red to green), Bonde buys. The stop is at the low of the entry day. For shorts the same works in mirror form, and there Bonde enters this way almost always.

Jack Corsellis waits longer. After an earnings gap on huge volume with a strong candle, the stock builds a base: a flag, a pennant, a Darvas box. Entry is on a narrow candle, a shakeout tail or a reversal after a gap down at EMA 10 or EMA 21, with the stop below the candle's low. The entry itself is the same as in trigger-bar-base. This card adds the earnings context and Jack's own screen.

A person judges the catalyst. Platforms cannot classify news, so the event below is defined by a gap and volume proxy.

Why it might work

Bonde's explanation. Most people watch a stock on earnings day and then forget about it. The catalyst has not gone anywhere: if the news changes the company's prospects, the second move comes after a pause. On a delayed entry the trader already knows the catalyst, is watching the stock and can use a tight stop. On the short side the logic is the same: after a gap down, buyers decide the worst is priced in and keep buying for several days, and the real decline starts later. Bonde does not explain why the entry moment is specifically the move through yesterday's close.

Jack explains the base through the behavior of institutions. After a surprise they keep buying on pullbacks to EMA 10 and EMA 21, volume in the base declines, and contraction gives way to expansion.

Neither explanation has been tested, and neither video contains statistics.

Rules

Event and watch list (author, Finetiq proxy)

// author: a person judges the catalyst (earnings, guidance, agreement, approval)
// Finetiq: a daily-bar proxy for testing without news
EventDay = (Open / Close[1] - 1 >= 10% OR Volume > Highest(Volume, 250)[1])
           AND Volume >= 3 * Average(Volume, 50)[1]
Failed   = Close < Open                                   // author: HIMS opened and closed lower
           OR High - Close >= 0.5 * (High - Close[1])     // author: Reddit gave back the gain; Finetiq: half of the move given back
// Finetiq: thresholds 10%, 250 and 3 × 50 days

IF EventDay AND Failed
    WatchList.add(symbol, EventDate)
    WatchUntil = EventDate + 21 trading days              // author: at least a month; Finetiq: 21 days
// author: EP 9 million days also go here (volume above 9 million shares, price from $3),
// after which the stock pulled back. Test a variant without the Failed condition

Red-to-green entry (author, details Finetiq)

// intraday data, every day after EventDate until WatchUntil
PrevClose = Close of yesterday's regular session
ADRpct    = 100 * (Average(High / Low, 20) - 1)          // on daily bars

IF symbol in WatchList AND Open < PrevClose              // author: the day opened in the red
    BUY STOP at PrevClose + 0.01                         // author: the moment of the turn from red to green
    StopLvl = LowOfDay - 0.01                            // author: stop at the low of the entry day
// Finetiq: if the day opened in the green, there is no turn and no signal
// author: sometimes 2-3 attempts in one stock; Finetiq: no more than 3 entries per event
// Finetiq: skip stops wider than ADR% (adr-stop-gate), widen stops tighter than 1/4 ADR% to 1/4 ADR%

Size and exit (Finetiq, based on Bonde's general rules)

R      = EntryPrice - StopLvl                           // initial risk per share, fixed at entry
Shares = MIN(Equity * 0.5% / R, Equity * 25% / EntryPrice)   // Finetiq
// the author puts 60-80% of the account into such entries. That is a conviction call and cannot be turned into a formula
IF Close > EntryPrice THEN StopLvl = MAX(StopLvl, EntryPrice)  // author (general EP rule): breakeven
IF Close >= EntryPrice + 2 * R THEN StopLvl = MAX(StopLvl, Low[1])
// author: then below the low of each day; Finetiq: switched on after 2R

Short mirror (author, entry Finetiq)

// author: on the short side 99% of entries are delayed. A gap down on a bad catalyst (example: sales guidance −44%),
// then several days of bounce, a short "on the reversal". The author did not name a reversal rule

// Finetiq: mirror event, thresholds as for the long side
ShortEventDay = (Open / Close[1] - 1 <= -10% OR (Volume > Highest(Volume, 250)[1] AND Close < Close[1]))
                AND Volume >= 3 * Average(Volume, 50)[1]
IF ShortEventDay
    ShortWatchList.add(symbol, EventDate)
    WatchUntil = EventDate + 21 trading days              // author: at least a month; Finetiq: 21 days
// Finetiq: no Failed condition for the short side, the entry itself waits for the bounce after the gap

// every day after EventDate until WatchUntil
IF symbol in ShortWatchList AND Open > PrevClose          // Finetiq: mirror, green to red
    SELL SHORT STOP at PrevClose - 0.01
    StopLvl = HighOfDay + 0.01

Variant B. Jack Corsellis: a base after an earnings gap

// daily bars
GapDay = earnings date AND Open / Close[1] - 1 >= 3%      // author: earnings gap; Finetiq: 3% threshold
         AND Volume >= 3 * Average(Volume, 30)[1]          // author: volume many times the 30-day average; Finetiq: 3
         AND Close - Low >= 0.7 * (High - Low)            // author: strong candle, ideally an open near the low; Finetiq: 0.7
         AND RSLine = Highest(RSLine, 250)                 // author: RS line versus the S&P 500 at a 52-week high
// author, optional: earnings and revenue surprise, raised annual guidance

// base after the gap (author: flag, pennant, wedge, Darvas box, VCP, cup handle)
BarsSinceGap >= 5 AND BarsSinceGap <= 40                  // Finetiq; SEDG has a 36-bar base (author)
Lowest(Close, BarsSinceGap) >= Low(GapDay)                // Finetiq: the base does not close below the gap day

// entry candle at EMA 10 or EMA 21, at SMA 50 for large bases, volume below the 30-day average (author)
EntryBar = TriggerBar (trigger-bar-base) OR shakeout tail OR reversal after a gap down (gap-down-reversal)
EntryLvl = High(EntryBar) + 0.01
StopLvl  = Low(EntryBar) - 0.01
RiskPct  = (EntryLvl - StopLvl) / EntryLvl * 100          // ADRpct as in the red-to-green block
IF EntryBar = TriggerBar  AND RiskPct > 2/3 * ADRpct THEN skip   // author: 1/2-2/3 ADR%
IF EntryBar <> TriggerBar AND RiskPct > 1.5 * ADRpct THEN skip   // author: up to 1.5 ADR%
IF RiskPct > 5 THEN skip               // author: more than 4-5% needs a justification; Finetiq: not allowed
BUY STOP at EntryLvl                                      // author: buy stop limit

// trade management (author): 1/3 or 1/2 at 1R, part on a close below EMA 10, the rest below EMA 21
SELL 1/3 LIMIT at EntryPrice + (EntryPrice - StopLvl)
IF Close < EMA(Close, 10) THEN SELL 1/3 AT NEXT BAR OPEN
IF Close < EMA(Close, 21) THEN SELL remainder AT NEXT BAR OPEN
// author: bearish synchronicity, open near the high, close near the low, break of EMA 10 and 21
BearSync = Open >= High - 0.1 * (High - Low) AND Close <= Low + 0.1 * (High - Low)
           AND Close < EMA(Close, 10) AND Close < EMA(Close, 21)          // Finetiq: 10% of the range
IF BearSync THEN EXIT all AT NEXT BAR OPEN

// stock selection, a separate rule (author, US Post Gap Consolidation screen):
// market cap > $300 million, Close > 1, above EMA 21, SMA 50 and SMA 200,
// 50-day average volume > 100 000, ADR% > 2.5, 3-month gain > 50%, sorted by ADR%
// the screen does not look for the gap, the author finds it by eye; Finetiq: in the test the gap is found by the GapDay rule

Parameters

Parameter	Value	Source
Event	gap of 10% or more or record volume, volume of at least 3 × the 50-day average	Finetiq (author: catalyst)
Failed day 1	close below the open or half of the move given back	author, threshold Finetiq
Watch window	21 trading days	author (at least a month)
Entry	open in the red, move above yesterday's close	author
Stop	low of the entry day	author
Attempts per event	up to 3	author (2-3), limit Finetiq
Risk per trade	0.5%, cap of 25% of the account, stop 1/4-1 ADR%	Finetiq (author: 60-80% of the account)
Trade management	breakeven, then the day's low after 2R	author, threshold Finetiq
Short	gap down of 10% or more or record volume on a decline, then green to red after a bounce	author (idea), event and entry Finetiq
B: gap	3% or more on earnings, volume of at least 3 × the 30-day average, RS line at a high	author, numbers Finetiq
B: base	5-40 bars	Finetiq (author's example 36)
B: entry candle	trigger bar, tail, reversal after a gap	author
B: risk	1/2-2/3 ADR% or up to 1.5 ADR%, no more than 5%	author (fractions of ADR%), Finetiq (ban above 5%; for the author more than 4-5% needs a justification)
B: exit	1/3 at 1R, closes below EMA 10 and 21, bearish synchronicity	author, threshold Finetiq (10% of the range in BearSync)
B: screen	market cap from $300 million, ADR% from 2.5, 3-month gain from 50%	author

What to test

Red to green versus simple entries. On the same events: a move above yesterday's close, a buy at the open of the first red day, a buy on a close above yesterday's close. Bonde claims this entry gives the best risk-to-reward ratio. Compare the average trade in R.
Is the failed day 1 needed. Run with and without the Failed condition, that is, after any strong gap followed by a pullback. Bonde uses both cases.
Window and number of attempts. 5, 10, 21 and 42 days; 1 and 3 attempts. If almost all the profit comes in the first days, a long window only adds losing attempts.
Stop width. On a red-to-green day the day's low is often very close. Split trades by stop size in fractions of ADR% and check how many get stopped out on the same day.
Variant B by candle type. Trigger bar, shakeout tail and reversal after a gap down separately. Base length of 5-15 and 15-40 bars.
Event proxy. Earnings dates only versus a gap with volume and no news. If the proxy works noticeably worse, the edge lies in judging the catalyst.
Costs and the short side. Everyone sees yesterday's close, and a stop order at that level slips. For shorts, add the cost of borrowing shares, which can be high for stocks after bad news.

Platform notes

TradingView (Pine Script)

Yesterday's close on an intraday chart without lookahead: request.security(syminfo.tickerid, "D", close[1], lookahead = barmerge.lookahead_on). If the chart includes extended hours, "red" and "green" are measured from a different close.
Set the event date as an input or an array of dates. A gap scan across all stocks cannot be reproduced in a strategy, because it trades only the chart symbol.
A stop order in Pine stays active until canceled. Cancel it at the end of the day with strategy.cancel, otherwise it carries over to the next session, where yesterday's close is already different.
The entry and the day's low are often hit by the same candle. Without use_bar_magnifier = true or a lower timeframe, the sequence is unknown.

MultiCharts and TradeStation (EasyLanguage)

The daily series is added as a second data stream, and yesterday's close is Close of Data2 on the bars of the current day. Check the symbol's session settings: the close of the regular session and the close of extended hours differ.
Buy next bar at PrevClose + 0.01 stop lives for one bar. Send it on every bar as long as the day opened in the red and there is no position.
Time is the bar's closing time. Accumulate the day's low yourself from the first bar of the session.
Variant B is calculated on daily bars. The event list and the scan across many stocks are done in Portfolio Maestro or Portfolio Trader.

MetaTrader 5 (MQL5)

Brokers offer stock CFDs mostly on large companies. Small and mid caps, where such events happen most often, are missing.
A CFD's daily bar is built on server time. If the CFD is quoted during extended hours, yesterday's close and the move through it do not match the exchange.
There is no real volume, only tick volume. The volume-based event proxy and the 9 million share threshold cannot be reproduced.
A CFD short does not use borrowed shares but carries a swap. The result of the short side will differ from a short on the exchange.

Where the idea can break

There are no statistics. All of Bonde's examples (HIMS, Netflix, Reddit, Mobileye) were picked after the fact, and a size of 60-80% of the account is Bonde's personal conviction call.
A person judges the catalyst. Bonde considers an event "valid" based on a personal reading of the news, and our gap and volume proxy may pick entirely different events.
A move through yesterday's close happens often, including in stocks with no follow-through. A tight stop produces many small losses, and costs weigh heavily on them.
The one-month window after the event overlaps with the overall market move. In a falling market, delayed entries can fail across the board.
Variant B describes the base and the entry candle in the author's words, and our thresholds for base length and candle strength may select different charts.
Bonde's video was published on a channel that advertises a prop firm. Jack Corsellis's video is sponsored by MarketSmith and promotes Jack's paid screener.