Power Drops and Failed Bounces: Short Entries and Risk

Video: THE STRATEGY BEHIND MY BIGGEST DAY TRADING PROFITS
Creator: Oliver Velez Trading
Original Video: Watch the original video
Difficulty: Advanced

The Big Idea

Oliver Velez describes a short-side sequence: a powerful drop, a bounce that fails to recover the move, and a second decline. He calls versions of this drop–rally–drop structure a “whale” and illustrates taking profits on the first decline, then restoring short exposure during the bounce.

The pattern is a continuation hypothesis, not a law of motion. A sharp drop can be followed by a weak rebound, a full recovery, or a genuine reversal. The useful work is defining the original context, measuring the bounce, and deciding where the short thesis becomes wrong before adding exposure.

The creator’s claims about profit contribution, odds of failure, and shorting near moving averages are not established by a trade ledger in this transcript. This article separates his discretionary demonstration from a mechanical research approximation.

Read the Context Before the Red Bar

The main example uses a two-minute Meta chart. Velez first examines the 20- and 200-period moving averages, a relatively flat 200, an opening gap above the averages, and a nearby round number. He distinguishes opening near the averages from opening more extended above them and cautions against shorting too close to that support area.

The “zones” are visual categories here; the video does not give numerical boundaries. A chart’s apparent 45-degree slope also changes with zoom and scaling. If testing, measure moving-average change in price or volatility units instead of screen angle. The transcript does not establish the moving-average calculation type, so any SMA or EMA choice in code is a documented research choice.

A whole-dollar price may be a useful reference because participants can focus on round numbers. It is not guaranteed resistance just because price gaps upward into it, nor guaranteed support after a gap down. Record the level before watching the response and require your chosen confirmation.

The First Short Is an Intrabar Decision

Velez watches the opening candle initially rise, lose that green movement, and become sufficiently red. He enters while the candle is developing, with a stop above the high already observed at that moment. He compares the developing red movement with recent solid red bars to judge whether the weakness is meaningful.

That is not equivalent to waiting for the completed two-minute candle. Historical OHLC bars do not reveal every color change, bounce, or order of extremes inside the interval. A backtest cannot inspect the finished candle and then pretend its final high was already known at an earlier intrabar entry.

The source’s quoted stop prices vary, including apparent transcription errors. A consistent illustration of a $0.75 stop is a $322.25 protective level and a $321.50 short entry, before costs. A $322.50 stop with the same entry instead creates $1.00 of price risk. Confirm the actual price references rather than combining conflicting numbers.

The Bounce Trade Is a Separate Decision

After the initial decline, the source describes taking part of the profit near another round number and placing orders to restore short exposure around halfway back up the drop. Later he distinguishes an aggressive entry that cuts into a green bounce from a more conservative entry below that green candle.

The video’s ball-bounce analogy explains the creator’s intuition about a rebound after a powerful fall. Market prices are not governed by that physical analogy. A bounce can recover the origin of the drop, and short exposure must remain controlled when it does.

Scaling Out and Back In Changes the Risk

Realizing profit on one portion does not make the remaining or restored position risk-free. Track the outstanding shares, average entry, protective levels, and worst planned incremental loss. A realized gain is part of the day’s result, not permission to remove the stop.

For an original example, short 200 shares at $321.60 with a $322.25 stop. Nominal initial price risk is $130. Cover 100 shares at $320.10, realizing $150 before costs. The remaining 100 shares still have $65 of nominal risk to the original stop.

Now add 100 shares at $320.85 while retaining the $322.25 stop. The added shares contribute $140 of price risk. Combined remaining-position risk to that stop is $205, against $150 already realized, leaving a possible net day result of −$55 before fees and slippage. The profitable first leg does not eliminate the cost of being wrong on the bounce.

If both remaining lots are covered at $320.00, their profits are $160 and $85, bringing total gross profit to $395. This shows how two exposures can earn from overlapping price territory. It does not mean the underlying price moved twice as far or that the two legs are independent trades statistically.

Micro and Macro Patterns Require Different Evidence

The source distinguishes a super-micro sequence within one bar, a micro version across several bars, and a larger macro pattern. A completed red candle may contain an internal rebound invisible in OHLC data. Detecting that internal sequence requires appropriate lower-timeframe or tick information and matching execution assumptions.

The larger pattern gives more bars to define the bounce, but also creates more choices about anchors and entry timing. Keep each version separate in the journal. Do not combine early intrabar shorts and slower multi-bar confirmation entries into a single performance claim without identifying their differences.

Velez mentions covering near his “Fab 4” buy area, but this excerpt does not define that area’s full calculation. It is therefore not a complete mechanical exit rule. Define a target or management model independently rather than inventing his omitted settings.

Pine Script: A Confirmed-Bar Failed-Bounce Test

//@version=6
// Guide: https://candlechasers.com/power-drop-failed-bounce-short-risk/
// Confirmed-bar continuation proxy, not Oliver Velez's intrabar entries.
strategy("CandleChasers - Power drop failed bounce", overlay=true,
     initial_capital=25000, pyramiding=0,
     commission_type=strategy.commission.percent, commission_value=0.01,
     slippage=2, margin_short=100, margin_long=100,
     calc_on_order_fills=false, calc_on_every_tick=false, process_orders_on_close=false)
shares = input.int(1, "Fixed shares", minval=1)
bodyLength = input.int(20, "Prior body average", minval=5)
powerMultiple = input.float(1.5, "Power body / prior average", minval=0.1)
minBodyFraction = input.float(0.6, "Power body / total range", minval=0.1, maxval=1)
maxRetrace = input.float(0.75, "Maximum bounce fraction of power range", minval=0.1, maxval=1)
maxWait = input.int(6, "Pattern expires after bars", minval=2)
rr = input.float(2.0, "Signal-price reward/risk", minval=0.1)
useMA = input.bool(false, "Require extension above 20 and flat 200 at power bar")
extension = input.float(0.5, "Minimum distance above 20 / ATR", minval=0)
flatness = input.float(0.2, "Maximum 200 change over 10 bars / ATR", minval=0)
ma20 = ta.sma(close, 20)
ma200 = ta.sma(close, 200)
atr = ta.atr(14)
body = math.abs(close - open)
averageBody = ta.sma(body, bodyLength)
barRange = math.max(high - low, syminfo.mintick)
power = close < open and body >= powerMultiple * averageBody[1] and body / barRange >= minBodyFraction
context = not useMA or (open > ma20[1] and open > ma200[1] and open - ma20[1] >= extension * atr[1] and math.abs(ma200[1] - ma200[11]) <= flatness * atr[1])
nyMinute = hour(time, "America/New_York") * 60 + minute(time, "America/New_York")
window = nyMinute >= 570 and nyMinute < 660
valid = chart.is_standard and timeframe.isminutes
if barstate.isfirst and not valid
runtime.error("Use standard candles on a minute-based timeframe.")
var int stage = 0
var int powerBar = na
var int bounceBar = na
var float powerHigh = na
var float powerLow = na
var float bounceHigh = na
var float bounceLow = na
var float frozenStop = na
var float frozenTarget = na
bool entrySignal = false
if valid and barstate.isconfirmed
    if not window or nyMinute == 570
        stage := 0
    if strategy.position_size == 0 and window
        if stage > 0 and (bar_index - powerBar > maxWait or high >= powerHigh)
            stage := 0
        if stage == 2 and bar_index > bounceBar
            if high > bounceHigh
                stage := 0
            else if close < bounceLow and close < open
                frozenStop := bounceHigh + syminfo.mintick
                frozenTarget := close - rr * (frozenStop - close)
                strategy.entry("Short", strategy.short, qty=shares)
                strategy.exit("Exit", "Short", stop=frozenStop, limit=frozenTarget)
                entrySignal := true
                stage := 0
        if stage == 1 and bar_index > powerBar and close > open and high > powerLow and high <= powerLow + maxRetrace * (powerHigh - powerLow)
            bounceHigh := high
            bounceLow := low
            bounceBar := bar_index
            stage := 2
        if stage == 0 and not entrySignal and power and context
            powerHigh := high
            powerLow := low
            powerBar := bar_index
            stage := 1
    if nyMinute == 958
        strategy.close_all(comment="Regular session end", immediately=true)
if strategy.position_size < 0
    strategy.exit("Exit", "Short", stop=frozenStop, limit=frozenTarget)
plot(ma20, "20 SMA proxy", color.blue)
plot(ma200, "200 SMA proxy", color.yellow)
plot(strategy.position_size < 0 ? frozenStop : na, "Frozen stop", color.red, style=plot.style_linebr)
plot(strategy.position_size < 0 ? frozenTarget : na, "Frozen target", color.green, style=plot.style_linebr)
plotshape(entrySignal, title="Confirmed bounce failure", style=shape.triangledown, location=location.abovebar, color=color.red, size=size.tiny)

The strategy below tests only the later continuation idea on standard two-minute US equity charts. A large red body relative to the preceding body average arms the model. A later green candle must stay within a defined retracement fraction of that power candle. A subsequent red close beneath the green candle’s low generates a short order for the next available bar.

Research choices are explicit: the first qualifying green bar defines the bounce; a later bar exceeding its high cancels that candidate; the setup expires after a chosen number of bars; the stop is one tick above the bounce high; and the target is a fixed R multiple measured from the signal close. This is more restrictive than the source’s flexible bounce descriptions and does not implement his intrabar opening entry, partial profit-taking, or add-back sequence.

The optional moving-average context filter is OFF by default. When enabled, it approximates extension above a 20 SMA and a relatively flat 200 SMA at the power candle using prior confirmed values and ATR. It does not reproduce the source’s zone labels, verify a gap, select a round number, or establish the support context. Those are additional variables to review or define in separate experiments.

The script requires a two-minute standard chart, permits new setups and entries from 9:30 through 10:59 New York time, and prevents pyramiding. It submits fixed stop and target prices with the entry; next-open fills, gaps, and slippage can materially change the nominal R ratio. Fixed shares are not constant monetary-risk sizing.

State changes and signals wait for completed bars. There are no higher-timeframe requests, future pivots, backdated signals, or order-fill recalculations. Keep tick and order-fill recalculation disabled. This approach deliberately gives up the source’s early intrabar timing; using final candle information to reproduce that earlier timing would introduce hindsight.

Any remaining position is closed on the completed 15:58 two-minute bar with an immediate simulated close. That same-close fill is not guaranteed live. Holiday early closes and missing final bars are not modeled; exclude affected dates or add a proper session calendar before operational use. Borrow availability and short-sale execution restrictions are also not modeled. Code has been reviewed for chronological logic but has not been compiled or backtested in TradingView here. TradingView’s repainting documentation (paid link) explains why developing bar values and historical final values require different treatment.

Test the Continuation Hypothesis

1. Power definition: compare fixed body multiples while retaining the same bounce and exit rules. Do not choose a threshold only from attractive examples.
2. Bounce depth: compare shallow and deeper retracements, recording candidates that cancel before entry as well as filled trades.
3. Context: compare the pattern alone with a predefined extension and flatness filter. Review gap and round-number context as separately documented additions.
4. Execution model: compare confirmed-close failure with a separately implemented stop-entry break below a completed bounce candle. Include missed fills and realistic costs.

Measure net expectancy, average win and loss, drawdown, adverse excursion, and independent setup count. Test scaling out and adding back separately from a single-entry version: restored exposure changes both payoff and downside. Reserve later data for evaluation after choosing definitions.

Final Takeaway

The useful sequence is power drop, limited bounce, and a defined failure trigger. The opening short, the bounce entry, and the add-back are different decisions with different information and risk. Make each one explicit rather than assuming that a powerful red candle guarantees the next decline.

Risk Disclosure

This article and code are for education and research, not financial advice or a recommendation to trade. Short selling and leverage can create substantial losses; a short position can have theoretically unlimited loss. Stops do not guarantee execution prices. Backtests may omit borrow fees and availability, short-sale restrictions, spreads, slippage, liquidity, gaps, holidays, and discretionary decisions. Simulated fills and historical results do not guarantee live performance. Validate the code, instrument specifications, and exposure limits before considering live use.

CandleChasers independently interprets this source and is not affiliated with or endorsed by Oliver Velez Trading. Profit and probability claims belong to the creator; the examples, research definitions, and code here are CandleChasers’ own educational work.