OptionBench — options scanner suite

[Kim]

 

A comment regarding M:

As of 10PM Eastern on Macy's Investor's URL, the Earnings date is not confirmed.

https://www.macysinc.com/investors/events-presentations-and-investor-updates/events/default.aspx

 

[Romuald]

On 8/15/2026 at 11:00 PM, Canuck_Dave said:

@Romuald .really like the changes to the entry readings the scales help show now vs historical ..... are there plans to introduce a short strategy to the mix - would be interesting to see how say a put credit spread has fared over the last number of cycles ... thanks for all your work to keep tweaking (improving) the tool 

 

Glad the scales read better: that was the intent, and you're the first to say so.

On short strategies: yes, and closer than you'd think. Credit put and call spreads already exist in the platform as ETF screeners, but they're general screeners — not anchored to an event. What you're describing is different: selling premium into an earnings or macro event to collect the IV crush afterwards. That's been in my backlog since July under exactly that framing.

The reason it hasn't shipped is that it's the mirror image of everything currently in there. Every strategy in the pre-earnings and pre-events scanners exits before the announcement: they buy the ramp in implied vol and never touch the move itself. A credit spread held through the print is the opposite trade: defined risk, but the event is now in your result rather than out of it. Different entry window, different exit rule, different backtest.

Worth doing though, and the machinery for the backtest already exists in my code. If you have a view on where the short strike should sit — 15 delta, 20 delta — I'd take it, because that's the parameter I'd otherwise pick arbitrarily.

 

[Romuald]

 

What a Monday morning looks like on OptionBench

Someone asked what the workflow actually feels like rather than what the features are. Easiest way to answer is to walk through this morning.

You land on Today. No search, no setup — the Opportunity Map is already there, showing what's in its entry window. One radio button switches between setups opening today and setups opening this week. Tiles are shaded by win rate, and each one carries the ticker, the entry date and the historical return.

 

 

Markets are still closed as I write this, and the card says so — pre-earnings scanned this morning at 09:11, pre-events still on Friday's data. That's deliberate: you should always know which session you're looking at.

Macy's caught my eye, so I clicked the tile and the row opened.

 

 

The top strip is the backtested edge: win rate over 16 cycles, average and median return, the size of the move this name usually makes on earnings, and the exact rule the statistics were measured under — Δ±0.25, one week out, exit on the first close at or above +10%.

Underneath, two things worth separating. On the left, the position as it prices right now — strikes, debit, and the fact that it's 100% extrinsic, which means it needs a move or a vol bid or theta gets it. On the right, whether that entry is expensive: cost cheaper than 75% of past cycles at this same point, implied vol at the bottom of its own range. Both needles sit left of the median, which is the reading you want.

Below that, what the market is pricing against what the stock usually does: ±12.1% implied versus a ±6.3% average earnings move. Nearly twice the usual — which matters when you're buying.

The playbook is four steps, collapsed by default. Timing, execution, what the trade is actually buying, and how it exits. The third one is the part most people get wrong: the position closes before the announcement, so the earnings move never enters the result. You're buying the ramp in implied volatility, not the gap.

And the history. Every past cycle, value and relative value, with the current cycle overlaid. Green dots hit the target, red didn't.

It gets busy fast, so you can strip it to the average band and read the shape instead.

That's the whole loop: land, scan, open one row, decide. Free to look at — https://optionbench.com

Not investment advice. The tool describes what setups have done historically; your broker's chain is where you check what actually fills.

 

[Canuck_Dave]

6 hours ago, Romuald said:

Worth doing though, and the machinery for the backtest already exists in my code. If you have a view on where the short strike should sit — 15 delta, 20 delta — I'd take it, because that's the parameter I'd otherwise pick arbitrarily.

Based on what was being done in another service I was subscribed to backtests were best around a 20 delta spread .... -40/+20 or -30/+10 were usually the ones used in trades

[Kim]

Follow-up on my note about M last week: Macy's has now confirmed earnings for September 10 — about a week later than the estimated date that was floating around.

 

This is a clean illustration of why we don't enter pre-earnings positions before the company confirms. Anyone positioned for an early-September report got repriced: options expiring before the 10th lose their event premium outright, and longer-dated positions built for a Sep-2 IV ramp now sit through an extra week of decay waiting for an event that moved. Nothing in a backtest protects you from this — the historical cycles all 'knew' their dates; your live trade doesn't until the company speaks.

 

The exceptions we make are narrow and deliberate: structures where the date risk is partially hedged and the relative value is cheap enough to compensate — a cheap calendar, a far-dated diagonal. If you're making that exception, know you're making it.

[Romuald]

2 hours ago, Kim said:

Follow-up on my note about M last week: Macy's has now confirmed earnings for September 10 — about a week later than the estimated date that was floating around.

 

This is a clean illustration of why we don't enter pre-earnings positions before the company confirms. Anyone positioned for an early-September report got repriced: options expiring before the 10th lose their event premium outright, and longer-dated positions built for a Sep-2 IV ramp now sit through an extra week of decay waiting for an event that moved. Nothing in a backtest protects you from this — the historical cycles all 'knew' their dates; your live trade doesn't until the company speaks.

 

The exceptions we make are narrow and deliberate: structures where the date risk is partially hedged and the relative value is cheap enough to compensate — a cheap calendar, a far-dated diagonal. If you're making that exception, know you're making it.

 

The right principle, and one the scanner tries to enforce rather than leave to memory.

Every pre-earnings row carries a confirmation flag: when the provider hasn't marked the report time as confirmed by the company, the card shows an amber "unconfirmed" chip rather than presenting the date as settled. M has been flagged that way throughout.

Straight about the current state though: the card still shows September 2. Macy's put out the September 10 confirmation an hour ago and my calendar feed refreshes each morning, so the corrected date lands tomorrow at the earliest. The chip says don't trust this date, which is right, but the date itself is stale until the provider catches up. I'm watching how long that takes.

Every cycle in the sample knew its date; the live position doesn't. No amount of better data closes that asymmetry, and it belongs in the trader's head rather than in a win rate.

On the narrow exceptions — cheap calendars, far-dated diagonals — those cards now show the executable debit next to the mid and the spread on each leg, after Yowster flagged the gap. If someone's taking date risk deliberately for cheap relative value, they should at least see what the fill really costs first.