Interview noteThis page restates a public interview in new words. Dollar results, win rates, and payouts are claims made by the speaker or the host. They are not this desk's record, and they are not a result to expect.
The firm puts two traders in one room for a live session. Gianluca Bruni is introduced as a Robbins World Cup champion who trades statistical expectancy and mechanical execution. Emre Gams is introduced as a former hedge-fund manager who trades options order flow and a discretionary read. They are not trading the same product or the same idea. Gianluca is trying to short Nasdaq momentum because the window is only a few hours and his main strategy will not print enough setups to beat variance. Emre is sitting in the S&P futures at a price where, he says, the most volume has changed hands, with an options position that was opened about a week before the cameras rolled. The host is explicit that the open gain on that options book, above $20,000, was not made in this session.
Gianluca's session plan
He does not ask why the market broke. His momentum model says that a session he can recognize as short has a positive expectancy only after a trigger, so the job is to see the trigger and not to explain the break. On the chart he marks a zone, and he does not care whether anyone calls it supply or demand. The break of that zone is what flips him. The value area is building inside the level, which he treats as a second hint. He identifies the two latest breakouts, including what had been the point of control of the prior long, and he says the session is now entirely below that level. From there he only wants momentum shorts.
The trigger is mechanical, and he refuses to describe it until the market prints it. The point of the morning, he says, is to beat variance, and variance is beaten with a lot of trades. His usual home is the fifteen-minute chart. Today he drops to the five-minute so more setups can appear in a short window. If the sample still does not show up, he has what he calls an emergency plan: a discretionary stab, which he does not want and will not force. The first condition is an extremely directional session. He reads structure in two layers, actual and peripheral. Both have to be pointing the same way before the picture is the one he prefers. Momentum, for this exercise, means at least ten minutes of downside, two five-minute candles, and he will take the same trigger on the S&P if the structure matches. One bullish candle and one bearish candle, alternating, and the setup never exists.
He quantifies the candle by how many ticks it travels. The threshold is something he fitted to a backtest and to an expectancy he is willing to live with. He is explicit that grabbing the single best threshold is how a test gets overfit. On the five-minute version they actually trade, the rule he shows is two candles of at least 60 ticks, both in the direction of a trend that the two structure layers agree on. Chop, where actual and peripheral disagree, is not the trade. The fifteen-minute chart, his real timeframe, has not confirmed yet. He takes the five-minute short anyway, ten contracts, because the session needs executions. He is not confident in the single fill. He says the trade has a positive expectancy if it is taken a thousand times, and that five or ten samples say nothing.
The risk is fixed before the target. The stop goes above the two signal candles, at the high of the first. The payoff is one and a half times that risk, every time. He sizes the contracts so the stop costs about $1,000. If the same pattern then prints on the fifteen-minute, he will add risk. Another pair of down candles while he is already in is another position, not a reason to hesitate. This is the version of the model he uses for prop accounts: he does not need a fat expectancy, he needs frequency. A candle that is merely red, or that is a plausible short on ordinary price action, does not count. He measures it. Under 60 ticks, it is not valid. He will not call a picture beautiful or ugly. It either respects the rule or it does not.
What "beat variance" means in his arithmetic
He illustrates the uncertainty of a small sample with a strategy he asks the room to treat as a 60 percent win rate at one-to-one, and only if the data survived the overfit checks. The formula he uses is one divided by the square root of the number of trades. On ten trades that is about 31 percent. Around a 60 percent win rate, ten outcomes can look like a 30 percent win rate or a 90 percent win rate. Luck and bad luck are still in charge. On a thousand trades the same formula is about 3 percent, and he puts the band at roughly 57 to 63 percent. A bad run still clears a one-to-one at 60 percent, so the edge, if it is real and if it does not decay, survives the noise. The host underlines the assumption: the win rate has to stay the win rate. Gianluca will not judge the model on one fill, or on ten. The smallest sample at which he says the uncertainty is low enough to mean something is about a hundred trades. A morning cannot produce a hundred. That is why a single stop does not interest him, and why he will take a slightly worse setup later in the session rather than sit in a sample of one.
He came to this from the other direction. He started in 2020, when what he was taught was to feel the market. Discretionary trading, he says, lost him a lot of money and his composure, because the feeling changed with the day. An angry morning was a different intuition. He does not believe an edge can live only in that intuition, and he thinks discretionary traders are often following a mechanical path they have not written down.
The backtest he is willing to show
The model on the screen is not his main strategy. It is built for a prop account, where the advantage he names is buying power. Frequency is the expectancy, in his framing. A strategy with a 0.2 expectancy and three trades a week can lose, inside a prop risk model, to a lower expectancy that fires six times a day. The numbers he reads off: expectancy of 0.12R, which he calls low on purpose. A luck simulation on the same data topped out around 40R. Their result was about 50R, and that gap is what he treats as the edge having cleared luck. Net return on the sheet is 59.8 percent at 1 percent risk, and he immediately discounts it. There are not enough samples for that figure to be realistic. What he does want the room to see is the variance test. The chance of still being in a loss after 100 trades is about 16 percent. After 406 trades it is about 2.3 percent. He does not think this version beats a benchmark on a risk-adjusted basis, and he would not run it on a personal account if the prop structure did not exist. With the risk model he built around it, he says it beats the prop constraints. That is the only claim he makes for it.
How the morning actually trades
The first short stops out. A level he was using then breaks, and he flips. The new trade is a long, toward a marked zone, and he is willing to take it against the value area. He calls it uncommon and still positive in expectancy, closer to a 55 percent coin than a 60 percent one, at one and a half times risk. Tails can still land seven times in a row. Over a hundred flips the weighting shows. He is late, takes the same contract count he had sized, and parks the stop under the low of the first signal candle.
He will add only if the next candle is also valid, meaning 60 ticks or more, and the stop for the whole position moves under that new structure. He calls the add pyramiding. The ticket reaches 22 contracts because the stop distance changed, not because he abandoned the dollar risk. With a stop near $2,000 the one-and-a-half target is about $3,000. They are still down about $900 from the first stop, so a full target would leave the session up about $2,100. The fifteen-minute still needs its own fifteen minutes. He would take that setup too, even into the zone he had marked, unless price has already reached the value-area high and printed a short trigger. The long existed only to travel to that level. A short trigger there closes the long and reverses. A single large red candle that does not meet the two-candle rule does not. He will sit near the target and let it work, because the coin on this trade is still the coin he bought. Two valid down candles would mean this coin is no longer rigged, the open gain is incidental, and the next edge is the short.
They never get the two consecutive candles. The long runs. After about three hours the target fills. He has made about $3,000 on the winner and about $2,000 on the session after the stop. The host's comparison is the one the video is built to provoke: Emre's $20,000 was largely off camera and already open, Gianluca's roughly $2,100 was taken under the rules while the room watched. Gianluca does not treat the comparison as the test.
Emre's book, and why he will not predict the snap
Emre's disagreement is immediate and narrow. He is not bearish just because Gianluca is. He has no directional bias at the inflection. He was short and he was long on the opening drive, inside a tight range, and now he is flat. Trading the next few seconds, he says, is how you get chopped. Something will snap. Front-running the snap is how a failed break runs you over.
The options position is a bet that price stays inside a range. While it does, the position is near its peak gain, and they are in the middle of it. The range he names on the S&P is 7870 down to 7770, a hundred points. He calls that a comparatively safe distance. He chose it by looking for where options look mispriced, the spot where he thinks institutional buyers will pay an inflated premium for protection. The call side, he says, has less of that demand, but he still wants it in the structure: if he is selling calls he wants the market to stay below the strike, and if the market instead leaks lower he wants that call side as protection. As price leaves the middle of the range he starts hedging with futures. He calls that gamma scalping. The futures are the tactical book. The options are the book.
On the execution screen he reads aggressive sellers in one color and aggressive buyers in another. Buyers lifting are being absorbed. Sellers pushing are meeting demand. He will not guess which side wins. He also has gamma-exposure levels. One of them, around 7832, lines up with a thin part of the overnight profile, roughly 7830 to 7836, where little business has been done. He expects the market to trade into that thin spot and then to find out whether the sellers who drove the overnight session will defend it. A level that is only a magnet is not yet a trade. He is a seller against it, not in front of it, and only if the defense shows.
The first hour is the initial balance. A wide one tells him the positioning is already large enough that a momentum chase is careless. That is the mirror of Gianluca's requirement. Gianluca needs the directional ten minutes or he has nothing. Emre is being paid, by the options, for the market to stay balanced. He puts a number on how often he thinks the other regime shows up. On a plain bell curve, most of the time price stays in balance. The extremely directional residue is about 30 to 35 percent of the time. He does not need to catch all of it. He says a correct read on about 30 percent of those breaks is enough for the returns to compound, and that needing to be perfect is how the "both books at once" idea fails.
He will not use a futures-style stop on the options. The options, as he has them on, make more when the market drifts gently higher than when it drifts gently lower, so his futures lean is more often short, as a hedge against that bullish lift. The question he asks is how much of the options gain he is willing to give back, not where a price line should sit. Futures are a second product. If the minor product starts consuming the margin of the main one, the business is wrong even if the futures trade was clever. He is down about $300 on the futures short while he says this, and he treats that as noise next to the options book. He will scale the short from one mini to two when the breakdown starts. He will not let it become the portfolio.
His reason for refusing a human delta-one fund, even one with a strong track record, is the same reason Gianluca abandoned discretion. Next year the gut may be a different gut. He mentions a friend running an algorithmic fund who did about 100 percent, or about 80 percent after fees, with a peak-to-trough drawdown well into the double digits. The return is real in the telling. The drawdown is outside Emre's tolerance. The shape he wants is steady systematic return, and then a slice of those gains spent hunting fragile, explosive edges, most of which will not work. Futures, in his account, can turn $20,000 into a few hundred thousand and will also lose $20,000 many times on the way. The expectancy can be positive and still be a bad business.
What the session is for
The two books only look opposed. Gianluca will take a 0.12R prop model a hundred times because the uncertainty band, not the last candle, is the object. Emre will sit inside a hundred-point options range and hedge the break he refuses to predict, because most sessions never leave balance and the ones that do are allowed to pay him only if the hedge is sized off the options gain, not off a hope. Neither of them is interested in a one-day scoreboard. Emre says a trading competition is one data point and is not a measure of skill, which is the same arithmetic Gianluca just did with one over the square root of ten. The morning ends with Gianluca at his target and Emre's older options gain still the larger number on the screen. The useful record is the rule each of them was willing to say out loud before the candle closed.