Leaderboard

Update — at the open Markets opened, and here's AMD live: the strangle re-priced from $38.69 to $37.79, AMD gapped up to $512, and entry IV eased from ~87% to 84.6%. Marginally cheaper — but the verdict didn't budge: still "Rich entry," still richer than 100% of past cycles, still flagged for crush risk. On the charts, the live-value diamonds sit at the very top of every panel — value, relative value, and IV. Translation: the setup got a little bit cheaper, not cheap. It never dropped into its normal P25–P75 band, so the disciplined read is unchanged. Personnally I would pass, and let the backtest stay a backtest for this cycle. That's the tool doing its job: a 95% historical win rate is only worth having if you don't overpay to get in, from my POV. Romuald - OptionBench

An 80% Win Rate on One of the Most Expensive Entries in the Ticker's History. Would You Take It? Here's a trade the OptionBench pre-earnings scanner flagged, and a question worth sitting with before you read the answer. [CAPTURE 1 — Win Rate 80% / Average Return +21% (median +22% · consistency 0.63) / Earnings Move avg ±3.0% implied ±3.1% / Next Earnings 2026-07-16] TSM reports earnings on July 16. The scanner flags a Long Strangle entered fourteen trading days out — on Thursday, June 25 — with Δ ≈ 0.25 on each side, expiring one week after earnings, taking profit at +10%. The headline numbers look strong: an 80% win rate over 20 past cycles, a +21% average return, a +22% median. Then you look closer, and two things should make you hesitate. First, the strangle is expensive. Its cost as a share of the underlying — the entry's relative value — sits at 4.5%, against a historical average of about 3.1% at this point in the cycle. Second, implied volatility is elevated: around 52% at entry, climbing toward 59%, well above the ~37% this name usually carries here. Every instinct trained on "buy low, sell high" says the same thing: it's expensive, IV is high, wait for it to come back down. And here's the part that should really give you pause — this entry isn't just above average. It ranks among the most expensive entries the scanner has on record for TSM at this point in the cycle, with implied volatility near the very top of its historical range. So: one of the most expensive entries on record, IV at its highs, and every reflex telling you to pass. Would you take the trade? The instinct that costs you the trade Most traders pass here, and their reasoning feels airtight: if it's the priciest it's ever been, mean reversion says it gets cheaper, so I'm overpaying. That reasoning conflates two different things — the price of the option and its expected value. They are not the same axis, and treating them as one is the single most common way traders talk themselves out of good pre-earnings trades. Let me show you why, with this exact trade, using numbers you can pull from the scanner yourself. "Expensive" is not the same as "overpriced" The first question a rigorous trader should ask isn't "is this expensive?" It's "is the market pricing the move correctly?" [CAPTURE 2 — Relative Value « is this trade expensive? », tooltip T-14 : 2026-06-25, Value 4.50%] The scanner answers this directly. Two figures sit side by side in the banner: Implied move at entry: ±3.1% — the move the strangle prices in, averaged across the 20 past cycles at this entry point. Realized earnings move: ±3.0% — the actual average move TSM has made on earnings, same 20 cycles. These are essentially equal. The market is not systematically underpricing TSM's earnings move, and it isn't wildly overpricing it either. The strangle is priced fairly relative to what the stock actually does. So if the option is fairly priced on the move, where does an 80% win rate come from? If the market gets the move right, the edge can't be coming from a mispriced move. It isn't. And that's the whole point. Where the edge actually lives The edge in this trade is not a bet that TSM moves more than the options imply. It's the capture of IV expansion into the event. [CAPTURE 3 — IV % : IV rising from ~52% (T-14) to ~59% (T-9), above historical band ~37%] You enter at T-14. The plan is a good-till-cancelled order to take profit at +10%, exiting before earnings and the volatility crush. And if the +10% never triggers? You still exit before the print — the position is closed at T-2 (for a before-market report) or T-1 (for an after-market one), no exceptions. You never carry the trade through earnings. That's the key risk control: this strategy has no exposure to the earnings gap itself, win or lose. In the window you hold the position, the option premium tends to inflate as the market crowds into the event. You're not holding through the print and hoping for a big move. You're buying anticipation and selling it a few days later, richer. [CAPTURE 4 — Value $, tooltip T-11 (2026-06-30) : Value $26.91, Return +37.5%, Rank 1/20] On the current cycle, that's exactly what played out. Entered at T-14 near $19.57, the strangle was marked at +37.5% by the close of T-11 (June 30) — and earnings hadn't even happened yet. Now, you wouldn't have pocketed +37.5%: your +10% GTC would have filled well before that, somewhere in the intraday tape as the position ran up. The realized gain is +10%, the take-profit. But that +37.5% end-of-day mark is the evidence that the target was hit comfortably, and early. The move you'd have been "waiting to see" is irrelevant to this trade, because you're out before it happens. This reframes "expensive" and "high IV" completely. They aren't warning signs to avoid. They're the fuel. A cheap, low-IV strangle would have far less premium to expand. You want the anticipation. You're selling it, not buying into it. (One honest note for the careful reader, because it matters: the "realized move" measures the one-day earnings gap, while the "implied move" is priced by an option that lives a full week past the event. They're not a perfect apples-to-apples comparison — the option captures the gap plus residual post-earnings vol. If anything, that makes the "fairly priced" read conservative. I'd rather flag it than have someone catch it and assume I was hiding it.) What should have driven the decision Not the price. Not the IV level. The conditional statistics — the numbers the scanner exists to surface: Win rate: 80%, conditional on entering at T-14, over 20 cycles — meaning 80% of those cycles hit the +10% take-profit before the event. (It's a TP-hit rate, not a vaguer "was it green" rate; worth being precise, since it's the number the setup is actually built around.) Average return +21%, and — this is the one that kills the "it's an outlier" objection — median return +22%. The median sitting at or above the mean tells you the average isn't propped up by one lucky cycle. The typical outcome is a strong winner. This is a regular edge, not a lottery ticket. That distinction is everything. An 80% win rate with a median of +22% is a fundamentally different animal from an 80% win rate where the average is dragged up by a single monster cycle while everything else limps. The scanner shows you which one you're looking at, and here it's the good kind. The honesty that makes this defensible I'm not going to pretend this is a free lunch, because it isn't, and the scanner won't let me pretend. There's a consistency ratio of 0.63 on this trade — mean return divided by the standard deviation of per-cycle returns. It measures how regular the edge is, and 0.63 is moderate. The edge is real and it repeats, but there's genuine dispersion: the 20% of cycles that lose, lose meaningfully. This is a trade to size with control, not to back up the truck on. That's not a caveat buried at the bottom. It's the point. A tool that only ever tells you a trade is great is a marketing tool. A tool that hands you an 80% win rate and a 0.63 consistency ratio in the same breath is giving you what you need to size the position honestly. And one more line I won't blur: the current cycle is a single live data point, and a single data point proves nothing. That the take-profit was hit early this time is encouraging, not evidence. The case for entering doesn't rest on it. It rests on the 20-cycle conditional distribution: 80% of cycles hitting the +10% target, a +22% median across all of them. The live cycle is an illustration of that statistic playing out, not the proof of it. If you take one thing from this piece, take that sentence. The takeaway Don't confuse the price of an option with its expected value. They live on different axes. An option can be among the most expensive it has ever been — near the top of its own historical range, IV at its highs — and still be the right trade, because the price of the option and the odds of the trade are two separate questions. "Expensive" answers the first. The 80% win rate, the +22% median, and the 0.63 consistency ratio answer the second. That second set is what should drive the decision. The whole reason a tool like this earns its place is that it puts both answers on the same screen — so you're not passing on a good trade because an instinct trained on the wrong axis told you it looked expensive. Numbers in this piece are pulled directly from the OptionBench pre-earnings scanner (TSM, Long Strangle, T-14 entry on June 25, 2026, 20-cycle sample). Every figure — implied vs. realized move, median return, consistency ratio, entry relative value — is visible in the tool. Verify them yourself; that's the point. Beta-testing is ending by the end of July but you can still give your feedback and ask for signing in on the OptionBench Beta Web Site.

@Romuald these revisions are a great step forward keep them coming

From a Green Tile to a Trade Decision: Inside OptionBench's Pre-Earnings Workflow Most screeners hand you a list. OptionBench hands you a decision — and then argues with you about it. Here's how that works, walking through a real setup on AMD. Start on the map The Today page opens on an Opportunity Map: where you can see one tile per backtested pre-earnings setup for the current week (or week ahead it it is the week-e,d). Each tile carries the ticker, the entry day, and two numbers: historical win rate and average return. The colour encodes the win rate: the darker the green, the more often the setup has worked across past earnings cycles. A "Trades of the week" toggle narrows the grid to the days just ahead, and the caption never lets you forget the key caveat: backtested, not realized. The darkest tile in this view is AMD — Mon, Jul 20 — Win 95% · +22.7%. Tempting. So we click it. One click into the analysis Clicking a tile doesn't dump you at the top of a table, it drops you straight onto that exact row in the scanner, ticker already in focus. For AMD, the setup is a long strangle entered eleven business days before the Aug 4 report and exited the latest about a week after, with a +10% take-profit. The top strip is the historical, backtested edge: a 95% win rate (19 of 20 cycles), a +22.7% average return, a +18.3% median, and a typical earnings move of ±7.2% versus ±4.4% implied today. On the backtest alone, this looks like one of the strongest setups on the board. The part that keeps you honest Here's where OptionBench stops cheerleading. The right-hand verdict reads "Rich entry — RV richer than 100% of past cycles here," and the cheapness panel flags elevated IV/RV at entry — crush risk, edge reduced. In plain terms: the strangle is currently more expensive than at any comparable point in its own history, so buying it now would mean overpaying for volatility, exactly how a great backtest quietly turns into a mediocre fill. One more detail to notice: the badge says "Last close · EOD." Markets are closed right now — it's Monday morning (in France), and the NYSE opens in about six hours — so every live number on the page is the last end-of-day snapshot, not a tradeable quote. The rich-entry read is real, but it's a photograph from Friday's close. The right move isn't to trade it. It's to wait for the open and see whether the morning re-prices the strangle cheaper, or confirms it's still rich. Reading the cycle charts To judge that, we drop into the cycle charts (next Figure). Three stacked panels track the trade from T-11 to T-1 (business days before earnings): the strangle's dollar value, its relative value (RV% — "is this trade expensive?"), and its implied volatility. Every faint line is one past earnings cycle; the bold black line is the average; green dots mark the cycles that hit the +10% target, red dots the ones that didn't. It's a lot to take in at once — so a single click ("unselect all," keep Average) strips it down to the essentials: the average path, the outcome dots, and a shaded band. What the shaded band means That band is the P25–P75 range — the middle 50% of past cycles at each point in the run-up. At every T-x day, we take all the historical cycles and shade from the 25th percentile up to the 75th; the dotted line running through it is the median (P50). Think of it as the setup's normal range. When today's live reading sits inside the band, the trade is priced about as usual; below it, cheaper than history; above it, richer than history. To be continued — at the open So AMD is a beautiful backtest with a currently rich entry, frozen at Friday's close. The interesting moment is only a few hours away: when the market opens, we'll watch whether the strangle cheapens back down into its normal band — and only then decide whether it's worth putting on. That's the whole idea. The map surfaces the opportunity; the analysis tells you whether today is a good day to take it. The discipline is in the gap between the two. See you at the open. Romuald - OptionBench Backtested results are historical and are not realized returns. Nothing here is investment advice. Figures: (1) Today — Opportunity Map · (2) AMD analysis header · (3) Full cycle charts · (4) Simplified view with the P25–P75 band.

@Romuald no rush I'm sure there are other more pressing issues that need your attention ... good luck with the maintenance over the weekend

Quick heads-up for anyone using or checking out OptionBench this weekend: app.optionbench.com will be in scheduled maintenance from this Friday 10th of July afternoon (France Time) through the weekend, back to normal Monday. We're doing a planned infrastructure migration on the backend — consolidating the data layer so the scanners run faster and cleaner as we grow. Nothing's broken; we'd just rather take it offline than serve half-migrated data. If you hit a maintenance page over the weekend, that's why — everything will be back Monday. If you were mid-analysis on something and want a hand once we're back up, just ping me and I'll help you pick it back up. Thanks for your patience — doing the plumbing properly now so the tool holds up as more of you come on board. — Romuald, OptionBench.com

@Romuald thanks for sharing those thoughts .... I did consider both of them before jumping in .... the richness factor did concern me somewhat and I realize there may be a reduced edge due to that .... I also looked at the other cycles and well not as good as 8 cycles they were all still acceptable in my books as an aside is there anyway to pick more than one strategy in the scanner .. . it would be nice to be able to choose straddles and strangles together as they are closely related ... thanks for the words of thanks it has been a great ride thus far in being able to play a small role in the development of OB

@Romuald first test run was a success as BAC hit 10% target just after open .... that write up is awesome very worthy of a second or even third read ... thanks

Good morning Romauld, Great Analysis and discussion! Thank you. Sarang

@Romuald here we go wish me luck 😁

Building OptionBench with you, not just for you What three weeks of beta feedback changed When we opened the OptionBench beta a few weeks ago, the plan was simple: put the tool in front of real options traders and listen. Sixteen of you took us up on it. A handful have been relentless — sending detailed questions, catching things we missed, pushing back when something didn't add up. This post is a thank you to those of you who've been active, and a look at what your feedback actually changed. Because it changed quite a lot. The thread running through all of it is one idea: put yourself in the trader's seat. Not "here is some data" — but "here is what you need to see, at the moment you're deciding whether to take this trade." You questioned the live values — so we built entry context One of you noticed that the live value of a position on the chart didn't line up cleanly with the historical cycles plotted next to it. A small thing on the surface. But it pointed at a bigger gap: the chart showed you where past cycles went, but not where you stand right now relative to them. So we added it. When the market is open and you're inside the entry window, the live banner now tells you exactly how today's entry ranks against every past cycle at the same point in time: 💰 Cheaper than 50% of past cycles at this point — RV 2.1% 📊 IV lower than 88% of past cycles — 24.9% Hover the live marker on any sub-plot and you'll see its rank there too — Rank 2/8, 5/8, whatever it is — so you can tell at a glance whether you're getting in cheap, average, or rich versus the stock's own history. This is the part we're most happy with, because it's exactly what experienced traders already do by eye when they read a relative-value chart. We just made it explicit, and put a number on it. You asked how much the stock actually moves — so we measured it A recurring question: how much does this name typically move on earnings? It's the first thing you want to know before buying a straddle or a strangle into a report. We now compute the realized earnings move directly from each stock's history — the average and maximum one-day move of the underlying on its past earnings reactions, properly aligned to whether the company reports before the open or after the close. We cross-checked the numbers against two other websites on several names and they line up. It's shown as a plain historical fact — this is how the stock has moved — not a prediction. What you do with it is your call. You read the RV chart by eye — so we drew the distribution If you trade calendars or volatility setups on SteadyOptions, you already live in relative-value charts. The usual approach is to eyeball the average line and ask "am I above or below it?" That works, but it hides something: a single average line tells you nothing about how spread out the past cycles were. Being "below average" means very different things depending on whether the cycles were tightly clustered or all over the place. So each RV and IV sub-plot now shows the dispersion band — the middle 50% of past cycles (25th to 75th percentile) plus the median — at every point before the event: Now "cheap" isn't a guess. You can see the full range the stock has traded in at this point in its cycle, and exactly where today sits inside it. The common thread None of these are flashy. They won't promise you an edge or a win rate. What they do is the same quiet thing: make visible, at the moment you decide, what you used to have to estimate in your head. Where does today's entry rank against history? How much does this stock really move? How wide is the range I'm trading inside? Those are the questions a careful trader asks anyway. We're just trying to answer them on the same screen, while it still matters. That's what "putting ourselves in the trader's seat" means to us — and honestly, we've only been able to get close to it because you've been telling us, in detail, where we were getting it wrong. Keep it coming The beta is free until 31st of July, and the feedback loop is the whole point. If something looks off, if a number doesn't match what you expect, if there's a question the tool should answer but doesn't — tell us. The most useful changes so far have all started as a message from one of you. Thank you for building this with us. Romuald & The OptionBench team