← AlgoxFlow
All 30 closed positions, in order, with the catalyst and the dealer-gamma regime behind each session — not just what we traded, but why the tape allowed it. Every row is in the public track record, and every aggregate on this page is computed from the rows on this page. Add up the column and you will get the same numbers.
Monday's week-ahead post recorded three things at the close, live from the gamma map: SPY at 763.65, sitting below its 764 put wall. QQQ at 706.19 after taking the worst of the session at −1.02%. IWM at 297.94, below its 298 put wall. All three majors in negative net dealer gamma — hedging that amplifies moves rather than absorbing them — and not one of them printing a clean zero-gamma cross.
Against that structure, SPY ATM IV was 10% and the chain was pricing an expected move of ±0.4% for a full session. Max pain sat above spot on all three (767 / 714 / 300). And 47 of the week's earnings reports were stacked into Wednesday and Thursday, most of them after the close.
So the setup was: no floor from dealers, almost nothing priced for movement, magnets overhead, and the year's biggest single report in the middle of it. What that produces is not a trend — it is a week that pays attention to which day you are in. Which is exactly how it traded.
Catalyst: none. Positioning week ahead of NVDA.
Regime: negative gamma, grinding at the walls, nothing priced. How we traded it: a positioning day, not a conviction day. With max pain overhead on all three majors and IV at 10%, the honest read was that the tape had no reason to pick a direction until Wednesday. So: five positions, all modest, three of them dated for later in the week rather than same-day. The one same-day index play — $QQQ 709C — was the loss, which is what buying an index call into negative gamma with no catalyst tends to get you. The $AMZN 265C came off flat and got logged as a scratch rather than dressed up either way.
| $MU 8/24 920C | +54% |
| $META 8/28 600C | +33% |
| $BE 8/28 240C | +25% |
| $AMZN 8/24 265C | 0% · scratch |
| $QQQ 8/24 709C | −33% |
Catalyst: the day before NVDA. Pre-earnings chop.
Regime: the classic dead zone — nobody wants a position into the print, so size leaves the tape and moves become fake. How we traded it: badly, and on purpose. Three same-day and next-day calls went against us in the same session and all three were cut at the plan: $MU −40%, $QQQ −39%, $TSLA −40%. Look at those three numbers together. They are the same number. That is not luck, it is a stop being honoured three times in a row on a day the tape gave nothing. The two that worked were both dated 8/28 — positions with room, not day trades. This is the day the week is actually built on: the −37% average loss is not a statistic, it is Tuesday.
| $META 8/28 590C | +30% |
| $MRVL 8/28 260C | +28% |
| $MU 8/26 1030C | −40% |
| $TSLA 8/26 360C | −40% |
| $QQQ 8/25 716C | −39% |
Catalyst: NVDA earnings, after the close.
Regime: the whole complex front-running one report. Semis dragged everything with them into the print, and in negative gamma that pull gets amplified instead of damped — the same structure that made Tuesday chop made Wednesday run. How we traded it: one lane, all session. Every position is a semi or a mega-cap tech name pulled along by them, and we stopped trying to be clever about direction the moment the lane established itself. Seven for seven, which happens when the regime and the catalyst point the same way and you do not fight either. Note $MU twice at the same 965 strike — +57% and +25% — the second lot deliberately smaller and later, taken because the lane was still working rather than because a new signal appeared.
| $AMD 8/26 490C | +84% |
| $INTC 8/26 88C | +84% |
| $PLTR 8/28 180C | +65% |
| $MU 965C · first lot | +57% |
| $AAPL 8/26 312.5C | +45% |
| $META 8/26 580C | +31% |
| $MU 965C · second lot | +25% |
Catalyst: NVDA beat. Reported to have added roughly $435B in market value — and the average stock fell.
Regime: the narrowest tape of the week. One name absorbed the entire bid and the breadth underneath it was negative — an index that looks fine and a market that isn't. How we traded it: we sized down to three positions, and that was the decision of the week that nobody will notice. When breadth collapses, the number of things that can work collapses with it, so trade count has to follow conditions rather than habit. $MSFT 502.5C caught the software bid for +38%. The $SPCX 143C swing was banked at +30% after a week's hold — a position opened in a different regime and closed in this one. $GOOGL 345C −30% was the right thesis on the wrong horse, which on a day this narrow is the standard way to be wrong.
| $MSFT 8/28 502.5C | +38% |
| $SPCX 8/28 143C · swing, ~1wk hold | +30% |
| $GOOGL 8/28 345C | −30% |
Catalyst: the broadening. Everything Thursday left behind caught up.
Regime: the mirror image of Thursday. Breadth came back and the bid spread out across everything that had been ignored for a session — and because it was the last day of the week, the same negative gamma that had amplified Wednesday amplified this too. How we traded it: wide, and we let one run. Ten positions across mega-cap tech, semis and small caps, because when breadth broadens the correct response is more lines, not bigger ones. The $AMZN 260C is the trade of the month and the mechanics matter more than the number: the first lot was banked at +46%, and only then did the rest ride from $0.59 to $6.15 — +942%, the largest single result in the log's 838-trade history. It ran on house money. Nobody held a full position through that move, and any recap implying otherwise is selling something.
| $AMZN 8/28 260C · runner · $0.59 → $6.15 | +942% |
| $AAPL 320C | +74% |
| $AMZN 8/28 260C · first lot | +46% |
| $SNDK 1565C | +36% |
| $MU 960C | +34% |
| $MU 932.5C | +32% |
| $GOOGL 345C | +27% |
| $IWM 299C | +24% |
| $TSLA 355C | +24% |
| $CRWV 88C | −41% |
The headline is 79.3% and an average winner of +81.2%. Both are true and both are misleading on their own, because one position was 50% of everything the week made. Here is the week with and without it.
| Average winner · all 23 | +81.2% |
| Average winner · excluding the $AMZN runner | +42.1% |
| Average loser · all 6 | −37.2% |
| Average trade across all 30 | +54.8% |
| Average trade excluding the runner | +24.2% |
| Payoff ratio (avg win ÷ avg loss) | 2.19 : 1 |
| Win rate needed to break even at that payoff | 31.4% |
| Win rate actually run | 79.3% |
Take the runner out and the week is still good — +24.2% average across 29 positions — which is the part that matters. A book that only works when it catches a 10-bagger is a lottery ticket. A book that works at +24% and occasionally catches one is a system.
The other line worth sitting with: at a 2.19:1 payoff, breakeven is a 31.4% win rate. We ran 79.3%. That gap — not the +942% — is where the week's edge actually lived.
1. The runner was 4.2× all six losses combined. Every loss on the week sums to −223 percentage points. The $AMZN runner alone was +942. One correctly-sized winner covered the entire cost of being wrong six times, with room to spare — and it only did that because the first lot was already banked.
2. Tuesday's three losses were −40%, −40% and −39%. Nearly identical, in the same session, in three different names. That is the signature of a stop being obeyed rather than a market being kind. The losses on this page are capped by decision; the wins are not capped at all. That asymmetry is the strategy.
3. Thursday was the smallest day on purpose. Three positions on the day breadth was worst, ten on the day breadth was best. Trade count tracked conditions, not appetite. Most losing weeks are built the other way around — most trades on the worst day, because a bad tape feels like it owes you something.
Not counted in any number above, because they are not closed: $ARM 9/18 320C is at −43% and being held on a September expiry rather than cut on an August tape. Plus $BMNR and $CROX swings still working. And one piece of housekeeping stated rather than hidden: the $NVDA 8/28 220C still needs its final result logged — the tracker dropped it, and it will go into the record when it is reconciled, whichever way it landed.
A regime week. Monday said there was no dealer floor and nothing priced for movement, and the five sessions after that were not one market but four different ones: a positioning day, a dead zone, a one-lane melt-up, a breadth collapse, and a broadening. The trades that worked were the ones that matched the day they were in. The trades that didn't were mostly index calls bought on days the index had no reason to move.
And the shape is the lesson, not the win rate. Take three planned −40%s on Tuesday without flinching, size down when breadth dies on Thursday, then be wide enough on Friday that one position can run 900% on house money. Capped losses against uncapped winners. This week is the textbook version of it, and textbook versions are rare — which is worth saying out loud too.
Lifetime record through this week: 619W / 189L · 76.6% across 808 decided positions.
This page is the week after it happened. Members got these as they were taken, with the strike, the entry, the gamma read behind it and the exit — plus the flow desk, the live conviction board and the AI copilot.
See what members get →
Educational market analysis, not investment advice. Options carry a substantial risk of loss and can expire worthless.
Per-position percentages on the contract, not account returns. Past results do not predict future results.
Full public track record · All briefings · Free gamma map