LessonThis is a lesson restated in new words. Any figure in it is an example from the source, not this desk's record.
Candace Lao says that every payout she has taken since she entered the prop business came from two ICT models, and that those payouts are in the millions. The host introduces her with a verified two and a half million dollars in payouts and sits with her on a five-hundred-thousand-dollar account. She scalps the Nasdaq only. She watches the S&P because, in her account, it has been leading, and she does not put the trade on the S&P. The millions and the verified figure are the host's introduction and her claim, not a fresh audit of the tape.
Why it matters
She copied the models first and then changed the risk, the entry, and the target to what she can stand to lose. She found the fit by manual backtesting. A discretionary entry, she says, has to be tested by hand, because an algorithm would not take the trades she actually takes. The test is not one pass through ten years and then a rest. The sample has to be large, and it also has to stay current. Thousands of people, the host says, already know the same concepts. The filter is which sweep, which gap, which stop, and when the picture is not a trade.
The idea in plain words
The first model is turtle soup, aimed at external liquidity. On the New York morning of October sixth, the nine-thirty candle sweeps a relative equal high. Relative means smooth and close, not a perfect double top. She reads that shelf as a pool of stops. She does not sell the sweep. If London and the hours before New York have already been grinding higher, she expects a retracement, and the quick poke through the high is the fake she wants. Before nine-thirty she also wants sell-side liquidity built on the way up, because that is the target under the market. The entry is an inverse fair value gap, and she will not see it on a slow chart. She runs a one-minute chart and, on a second screen, the fifteen-second and the one-second. In the example the gap is on the thirty-second. A textbook gap waits for a body close. Her version is aggressive. When price touches the forming gap and trades through the midpoint she is in, on a market order, inside a bracket. Beginners, she says, should wait.
The second model needs a market-structure shift and a retracement into a fair value gap. It is usually not the first leg. She calls it the second leg of distribution, and the stop is larger. On the October ninth New York example she is at the desk around nine, not at nine-thirty, because the nearest liquidity pool has to be known before the open. A swing high around nine-eighteen that sweeps minor buy-side would have been the cleaner turtle soup, and she skips it because it is still about fifteen minutes from the open. The nine-thirty candle instead sweeps a swing low and retraces. The first fair value gap of the New York session is, for her, the key level of that session. Her trigger inside the gap is the midpoint, not the edge. She will not sell the moment price ticks the low of the gap, and she also will not wait for the close the way she says beginners should.
How to use it
Speed is a filter. If the same turtle-soup picture takes many candles, she does not take it. A slow walk into the gap means the market wants more consolidation. The twenty-five-point stop is a standard, not a level. Twenty-four or twenty-six is not the point. If an order block or another gap would force the stop wider, she treats the entry as wrong and does not stretch the stop to fit it. The bracket's target order is fifty points, so the minimum payoff is two to one. She will scale. If a second contract has a larger target, the first can be taken nearer one and a half to one. She came to twenty-five and fifty from the London and New York range she actually trades, and from the backtest. If the Nasdaq were trading near forty thousand, she says the range would be larger and the stop would have to grow with it. The sweeps she cares about most are a choppy climb that builds sell orders underneath, relative equal highs, and session highs and lows. A fair value gap is also a target. If she is aiming at external liquidity and a large gap sits in the way, she wants the trade closed at the gap.
The second model is not a twenty-five-and-fifty bracket. Around forty points is the example. She will consider sixty or one hundred. She will not take a seventy-point stop just to hide above an order block. She would rather use the tighter stop, know she may be tagged, and accept that. Beginners, she says, should not learn that part yet. The first partial is the next swing low. The rest comes off at the new-week opening gap. She uses order blocks for stops, not as her ordinary target. Scaling is two contracts at the midpoint, then a wait. If a long wick says the bearish gap is real, she adds on the retrace, and the stop trails up to the order block. Otherwise the trail sits on a quarter level or the midpoint.
She puts her average win rate at forty percent, and she has also described it in the session as forty to forty-five. Six losers in ten is the base case. At that rate she will not take one and a half to one. Her data says that if she keeps taking it, she does not make money. The trade has to offer at least two to one or she passes. She does move the stop. She trails it. She does not widen it because a candle turned against her. London's kill zone, for her, starts at two in the morning Eastern. Asia's window, as she describes it, runs from about eight in the evening Eastern to midnight. The midnight opening price is a magnet. She will not enter a long before she has seen how price behaves there. A bearish higher timeframe with price in a deep discount is a reason to look for a long scalp back toward premium, not a reason to sell the low. If she is using only the Asia range, premium is the midpoint of that range. If the S&P has already tagged the Asia high and the Nasdaq's Asia high is still far away, she calls that bearish divergence and will not buy the Nasdaq into it. These days, she says, the S&P leads.
One winning trade ends her session. If the trades are losers, she has a dollar cap. On a one-hundred-fifty-thousand-dollar account the cap she names is three thousand dollars. Hit it, and she stops until nine-thirty. London starts at one contract and can scale to a second. New York starts at two and can scale two more. The first contract is the test. If that one contract stops, the idea was wrong and the damage is one contract. Beginners, she says, should not scale. She wants the trade assumed to be a loser before it is entered. She will not close because a bearish candle appeared. She makes one exception for circumstance: a trader with a single funded account who needs the payout for bills may have a reason to take the open gain. That is their constraint, not a new rule.
On the London session they are early for the two o'clock kill zone and late for the one she wanted. A long into a one-hour fair value gap fills late on a limit. She trails the stop to the swing low, about twenty points, and splits the targets: about thirty points on the first piece and about forty on the second. She does not like the climb. It is building sell-side liquidity on the way up. Price does return to the midnight opening, then retraces through her entry. She lets it. The stop is hit. Two contracts, a loss she puts at six hundred fifty-six dollars. The next long is not available just because she wants one. She wants the Asia low swept first. A long inside a back-and-forth London range, after a stopped long, would be revenge. She also will not trade an eight-thirty high-impact release. The examples she gives are CPI and NFP. She does not trade the midpoint of the range. Shorts belong in premium, longs in discount. She will not buy the breakout the host can see in hindsight. Being right about the direction still pays her nothing if the entry rule is not there. She treats a winning trade taken outside the rules as worse than the six hundred fifty-six dollars, because the win is what teaches her to break the rule again.
What usually goes wrong
The ordinary failure is treating a slow picture as the violent open the first model wants, stretching a stop to hide it behind an order block, or selling a deep discount because the higher timeframe is bearish. Another is deciding in advance that the session must be long or must be short, then holding the bias after the setup has disappeared. An inverse gap by itself is not the model. The model includes the narrative. If she cannot say what price is doing, she does nothing. The payout totals and the win rate remain her figures and the host's introduction.
Questions people ask
Why will she not take one and a half to one?
She puts the average win rate at forty percent, and also mentions forty to forty-five in the same session. At that rate, she says, her own data does not pay if she keeps accepting one and a half. Other people might make that payoff work. Her minimum is two to one. That is her claim about her sample, not an audit.
What is the twenty-five-point stop attached to?
The turtle-soup bracket, with a fifty-point target order. It is a standard, not a measured level. If the stop would have to widen past an order block or another gap, she treats the entry as early or late and skips it. The second model uses a wider, discretionary stop, with about forty points as the example.
Why did the live long lose six hundred fifty-six dollars?
The fill was late, the climb built sell-side liquidity, and she refused to flatten on a bearish candle because that would abandon the idea. Two contracts hit the trailed stop. She calls it a losing trade next to a win rate she has described as forty to forty-five percent, where the two-to-one is what pays for the losers. She then stops for the session.