Why iPhone swipe-typers can feel “blocked” — and why that matters for typing tests
If you’ve ever tried to swipe-type “United Ireland” and watched your iPhone morph it into “United Iceland,” you’ve run into Apple’s predictive text gray list. In November 2026, independent researcher Derik Kauffman surfaced a hidden, multilingual list—tens of thousands of words and phrases—that Apple’s iOS keyboard won’t autocomplete and often won’t allow via swipe-typing (QuickPath). Examples range from political jabs like “Dictator Trump” and “Sleepy Joe” to geopolitical terms such as “Tibetan independence” or phrases like “NSA spies.” Apple told reporters its keyboard is designed to avoid offering inappropriate or offensive suggestions, and that users can always type any word manually. (theguardian.com)
Kauffman published the decoded lists (drawn from iOS 27.0 keyboard data and Apple language downloads served in October 2026), showing coverage across many languages. Newsrooms summarized the scope as well over 100,000 affected terms across nearly 30 languages, with 4,513 phrases identified in English alone. (swipeblocked.com)
For swipe-typers, these blocks matter: tests by Reuters on a subset of phrases found they were consistently hard to enter by swiping or would not appear as suggestions, which can slow users, force backspaces, or inject wrong words mid-flow. That’s classic friction—exactly what a typing test measures. (straitstimes.com)
How a hidden list can skew “fair” results
Typing tests aim to compare humans, not keyboards. But when one platform quietly refuses to assist (or even hinders) specific sequences:
- Accuracy gaps emerge. Graylisted phrases can derail swipe paths, inserting neighbors with similar gesture shapes (e.g., “Iceland” instead of “Ireland”). (theguardian.com)
- Speed gaps widen. If you must peck a whole word that a competitor’s device completes instantly, your WPM suffers for reasons unrelated to skill. (ca.marketscreener.com)
- Content bias creeps in. If politically charged or culturally specific terms are more likely to be blocked, score differences can depend on vocabulary mix instead of ability. (theguardian.com)
Android’s default Gboard has its own preferences and learning behavior, but Google’s documentation emphasizes user control: you can add words to a personal dictionary, create text shortcuts, and even dismiss unwanted suggestions. There’s no public documentation of a comparable, systemwide gray list for Gboard, which means identical test passages may feel easier on Android by default. (support.google.com)
A practical blueprint: platform-fair drills for your typing test
To keep scores comparable across iPhone and Android—without watering down real-world difficulty—offer two complementary modes.
1) Cross-platform fairness mode (exclude graylisted terms)
- Build an “allowlist” of phrases for each supported language by removing all entries found in Kauffman’s Swipeblocked sets. Update regularly. (swipeblocked.com)
- When generating passages, exclude any n-grams that match graylisted continuations (e.g., don’t follow “United” with “Ireland” in iOS English drills). (swipeblocked.com)
- Label sessions “graylist-aware” so users know results are normalized across devices.
- Keep language parity. Swipeblocked includes per-language lists; use them so Spanish, Arabic, Hindi, etc., are equally fair. (swipeblocked.com)
2) Graylist stress test (opt-in)
An advanced, iPhone-specific mode that purposely uses graylisted sequences to surface friction, teach workarounds, and quantify impact.
- Design short, focused drills: e.g., triads of graylisted phrases separated by neutral text so users can feel the stumble points without overwhelming the test. (ithinkdiff.com)
- Track delta metrics: WPM drop versus a baseline passage, error bursts at blocked terms, and time-to-correction after a swipe miss.
- Offer coaching cards between rounds: “Try text replacement for this phrase” or “Train your personal dictionary by typing it fully twice.” (See tips below.)
- Be transparent. Explain that Apple’s keyboard filters suggestions and swipe paths for certain terms; link to the public lists and outside reporting. (swipeblocked.com)
Implementation checklist for test builders
- Source of truth: Pull lists from Swipeblocked (code/data noted on site) and cache with a version tag (e.g., “iOS 27.0, Oct 2026 language packs”). (swipeblocked.com)
- Fuzzy matching: Many entries are phrase continuations. Match multi-word n-grams and block likely swipe-confusions (suggest close neighbors for A/B testing). (macworld.com)
- Language detection: Pick the right per-language list based on the passage language; avoid mixing languages in one drill.
- Setting flags: Let users declare keyboard platform/model and whether Predictive Text is on or off; store this meta for analytics.
- Privacy: Never upload what users type. All graylist checks should run locally on your passage generator, not on keystrokes.
Actionable tips for test-takers (especially on iPhone)
- Use Text Replacement to “whitelist” sticky words. On iPhone, Settings → General → Keyboard → Text Replacement → +, then put the full word or phrase in “Phrase.” You can leave Shortcut blank or add a memorable trigger. This adds it to your personal dictionary. (support.apple.com)
- Train personal vocabulary. Apple notes the keyboard learns explicit language you use; deliberately typing a blocked term a few times (without relying on swipe) can help your device learn it for future contexts. (apple.com)
- Know when to toggle Predictive Text. Some iPhone users prefer to switch Predictive Text off for competitive tests to reduce interference, then switch it back on afterward. Settings → General → Keyboard → Predictive Text. (support.apple.com)
- On Android (Gboard), add words and shortcuts in Gboard → Dictionary → Personal dictionary. This can restore glide-typing flow for uncommon names or technical jargon. (support.google.com)
Keep results comparable—and honest
- Offer both modes side-by-side and badge the results (“Cross-platform” vs “Graylist stress”).
- Report the estimated graylist friction: show how much a user’s WPM and accuracy changed when graylisted sequences appeared.
- Publish your passage policy and data sources so competitors and researchers can reproduce your fairness settings. (swipeblocked.com)
Bottom line
Apple’s hidden gray list changes the feel—and sometimes the outcome—of swipe-typing. That doesn’t make iPhone users worse typists; it means their keyboard silently intervenes. By building graylist-aware drills and offering an opt-in stress test with practical coaching, typing test platforms can keep the competition fair while helping users master their everyday tools. (ca.marketscreener.com)