"Best practices we learned fitting 400 boilers" scored 74. It names 400 boilers
The empty-promise gate charged a line for its opening phrase and never read what followed. "Best practices we learned fitting 400 boilers" scored 74, "Learn more about the KX-9 recall" scored 74, and "Everything you need to know about the 14 Oct cutover" scored 70 — each names something a reader could not have guessed, and each was docked for the three words in front of it. They score 100, 100 and 96 now. The distinction the list was missing is worth having if you write subject lines for a living: a container phrase — best practices, the ultimate guide, tips and tricks, learn more, everything you need to know — names a genre whose contents can be entirely specific, so evidence inside it genuinely rescues it. A hype phrase — supercharge, to the next level, unlock your potential — claims value rather than promising content, and nothing can sit inside a claim, however many digits surround it. The five genuine empty promises are unmoved at 74, including "Best practices for handling callbacks", which opens on the same two words as the first line released. And the free subject-line app on yourorbit.team, which had been quietly scoring differently from the extension since April, now agrees with it.
What shipped
•Bug fix — orbit_score_subject_line reads the payload, not just the phrase. An empty-promise pattern used to fire on the match alone, so the model number in "Learn more about the KX-9 recall" counted for nothing against the two words in front of it. A container phrase is now cleared when a concrete anchor — a digit, a currency or percent symbol, or a personalisation token — sits in the text after the match. The phrase was never the signal; what follows it is.
•Two different faults were sharing one list, and only one of them can be rescued by evidence. Five phrases name a genre and are exemptible. The rest claim value, and a claim has no inside — "Supercharge your workflow with 5 easy wins" carries a number and stays docked, because the number is decorating the claim rather than filling it. All 16 adversarial filler lines carrying a digit or a currency symbol remain docked.
•The obvious version of this fix was built, measured and thrown away. Exempting any line that carried a concrete anchor released 7 of those 16, because a digit is not a referent — "10 tips and tricks to supercharge your growth" has one and is pure filler. Two conditions do the work instead: the anchor must sit after the phrase, since an anchor in front is the listicle counter that quantifies a promise rather than filling it, so "7 best practices for better results" is still docked; and it cannot be a non-referent, which is why a bare year, a multiplier and a count of the promised genre itself are erased before the test runs. "The ultimate guide to email deliverability in 2026" fails on that last rule and should.
•Behaviour change — the free subject-line app at yourorbit.team/apps/subject-line now scores what the extension scores. It is a separate implementation, last touched on 22 April, and nothing compared the two, so it drifted in silence for four months: the site was handing 100/sharp to lines the extension scores 66 and 74, and checking one line in both places gave you two answers. Content-emptiness and the empty-promise rules are ported and agree exactly. If you use the web app, it is stricter than it was, deliberately — you should know it moved under you.
•Orbit's jargon detector is deliberately not ported to the web app, and three lines still score differently because of it. It is 1,026 lines, and copying it would create a third implementation to drift. The widest remaining gap is 28 points on "Unlock the power of your marketing journey", which the extension charges for and the web app does not. All three are recorded with the gap measured on the day they were found, and that gap is allowed to close and never to widen.
•Still docked on purpose, and it will read as a bug the first time you hit it. "The latest news", "helpful information", "insights to help you" and "what we've been up to" are claims of recency or usefulness rather than containers, so nothing you put behind them clears the charge — "Important information about your 14 Oct cutover" names a real date and still scores 74. That is a known false positive, priced against the false negative of exempting four phrases that are empty in almost every line they appear in. It is the call, not an oversight, and it is one test bank away from being reversed if it turns out to be wrong.