Quick Answer
The -teen vs -ty distinction is one of the few pronunciation errors that changes the actual information an interviewer receives. 'Thirteen' and 'thirty' are different numbers. When you say 'we reduced latency by 15%' but the interviewer hears '50%', your credibility takes a hit — not because of your accent, but because of one unstressed syllable. The fix is specific: stress the second syllable for -teen words, the first syllable for -ty words.
Why this error is different from other pronunciation mistakes
Most pronunciation errors in interviews are noise — an interviewer mishears a word, context fills the gap, and the conversation continues. The -teen vs -ty error is different: it changes a number, and numbers in technical interviews are load-bearing. Your impact metrics, your system's scale, your team size, your latency targets — these are the facts an interviewer uses to judge the scope of your experience.
If you say you managed a team of 'fourteen' engineers but the interviewer hears 'forty', one of two things happens: they either ask a confused follow-up question (which breaks your flow), or they silently recalibrate their picture of your seniority upward — and then feel misled when your answer doesn't match. Neither outcome helps you. This is not an accent problem. It is a stress-pattern problem, and stress patterns are learnable.
Interview tip
Before drilling pronunciation, check whether your metrics are actually in your answer at all. Many non-native speakers under-quantify their work in English — not because they lack the numbers, but because pulling a figure into a sentence mid-answer feels risky. If that's your real obstacle, Mockly's guide on pronouncing numbers and metrics walks through both the language and the confidence piece together.
The guide worth reading before or alongside this one is pronouncing numbers and metrics — it covers the full landscape of number pronunciation in interview answers, including percentages, decimals, and large figures, so this article can stay focused on the -teen vs -ty distinction specifically.
The rule: stress position is everything
Insight
For -teen numbers, stress the second syllable (thir-TEEN). For -ty numbers, stress the first syllable (THIR-ty). That single difference is what separates the two words for a native English listener.
What is syllable stress?
Syllable stress means saying one part of a word louder, longer, and at a slightly higher pitch than the others — in English, stressed syllables carry the meaning, and unstressed ones are often reduced or swallowed.
English listeners use stress position as their primary signal for distinguishing -teen from -ty. The vowel quality in the final syllable also differs — the -teen ending has a long 'ee' sound that is held; the -ty ending is short and often reduced to something closer to 'tee' or even 'dee' in fast speech. But stress is the dominant cue. Get the stress right and the vowel difference follows naturally.
| Number | Correct stress | How to feel it | Common error |
|---|---|---|---|
| 13 — thirteen | thir-TEEN (stress on 2nd) | The 'teen' is longer and higher in pitch | THIR-teen (sounds like thirty to a listener) |
| 14 — fourteen | four-TEEN | Hold the 'teen', drop the 'four' | FOUR-teen (sounds like forty) |
| 15 — fifteen | fif-TEEN | The 'teen' rings out clearly | FIF-teen (sounds like fifty) |
| 16 — sixteen | six-TEEN | Stress falls on the second half | SIX-teen (sounds like sixty) |
| 17 — seventeen | sev-en-TEEN | Three syllables; stress is on the last | SEV-en-teen (sounds like seventy) |
| 18 — eighteen | eigh-TEEN | Hold the 'teen' | EIGH-teen (sounds like eighty) |
| 19 — nineteen | nine-TEEN | Both syllables are strong, but 'teen' is higher | NINE-teen (sounds like ninety) |
The pattern is consistent: -teen = stress last syllable; -ty = stress first syllable.
Watch Out
There is one genuine exception to watch for: when a -teen number comes directly before a noun, some native speakers shift the stress forward — 'THIR-teen engineers' rather than 'thir-TEEN engineers' — to avoid two stressed syllables colliding. This is called stress shift, and it is normal. In isolation or at the end of a phrase, always stress the -teen.
Where it breaks in real interview answers
The confusion almost never happens when you say a number in isolation. It happens when you are mid-sentence, thinking about what comes next, and the number is just a detail you are passing through. That is exactly when stress collapses onto the default (first-syllable) pattern — and 'fifteen' becomes 'fifty' in the listener's ear.
The highest-risk moments in a technical interview
- Quoting latency or response-time targets: 'We kept p99 latency under 15 milliseconds' — if 'fifteen' sounds like 'fifty', the number is wrong by a factor of more than three.
- Describing system scale: 'The pipeline processed around 13 million events per day' — 'thirteen' vs 'thirty' is a 2.3× difference in scale.
- Stating team or headcount: 'I led a team of 14 engineers' — 'fourteen' vs 'forty' changes the seniority picture completely.
- Giving percentage improvements: 'We reduced error rate by 18%' — 'eighteen' vs 'eighty' is the difference between a modest win and an implausible claim.
- Mentioning time periods: 'We shipped this in 16 weeks' — 'sixteen' vs 'sixty' is the difference between a quarter and over a year.
- Naming SLA or uptime targets: '99.19% uptime' — the '19' at the end is especially vulnerable when you are reading a number aloud quickly.
Key takeaway
The numbers most likely to be confused are 13/30, 14/40, 15/50, 18/80, and 19/90 — because the vowel quality in the unstressed syllable is closest in these pairs. Drill these five pairs specifically.
There is also a listening risk that goes in the other direction. When the interviewer asks you a question containing a number — 'How would you design a system for 40 million users?' — you need to be sure you heard -ty and not -teen. If you are uncertain, it is always correct to confirm: 'Just to make sure I have the right scale — that's forty million, with an F?' Native speakers do this too, and it reads as careful, not weak.
Which first languages make this harder
The -teen vs -ty error is not equally distributed across all speakers. It is most persistent in speakers whose first language does not use lexical stress — where syllable prominence is not the main signal that distinguishes words. If your language marks meaning through tone, vowel length, or syllable timing rather than through stress placement, your ear has genuinely not been trained to weight stress as a primary cue.
| First language group | Why this is harder | What to focus on |
|---|---|---|
| Mandarin / Cantonese | Tone languages: pitch signals word meaning, not stress. The stress-shift pattern in English is not a native cue. | Consciously exaggerate the length of the -teen syllable. Make it feel too long — it will sound right to a listener. |
| Japanese / Korean | Mora-timed or pitch-accent languages: syllables tend toward equal weight. Reducing the first syllable of -teen words feels unnatural. | Practice dropping the first syllable almost entirely: 'thuh-TEEN', not 'THIR-teen'. |
| Hindi / Bengali / Telugu | Stress exists in these languages but the patterns differ from English. The -ty ending can sound like -teen because the final vowel is lengthened. | Focus on shortening and reducing the final syllable of -ty words: 'THIR-dee', not 'THIR-tee'. |
| Arabic | Stress is phonemic in Arabic but the English -teen/-ty distinction doesn't map onto a familiar Arabic contrast. Both endings can collapse to similar sounds. | Record yourself saying pairs aloud and compare to a reference. The difference is audible on playback even when it isn't felt during production. |
| Russian / Polish / Ukrainian | Stress is strong and variable in Slavic languages, so the concept is familiar — but the specific English patterns for -teen vs -ty need to be learned explicitly. | The rule is learnable quickly. Focus on the five highest-risk pairs (13/30, 14/40, 15/50, 18/80, 19/90) rather than all nine. |
| Spanish / Portuguese / Italian | Syllable-timed languages: each syllable gets roughly equal duration. Reducing the unstressed syllable in -teen words feels like losing part of the word. | Allow the first syllable to be very short and quiet. The word is still complete — English just weights syllables unevenly. |
If your first language is stress-timed (German, Dutch, Russian), you already have the mechanism — you just need the specific English mapping. If it is syllable-timed or tone-based, you are learning a new perceptual category, which takes more deliberate practice.
Before and after: the same answer delivered poorly, then well
Here is the same interview answer — a response to 'Tell me about a performance improvement you led' — first with the stress error present, then with it corrected. The content is identical. The difference is only in how the numbers land.
❌ Before — stress error present
"We had a service that was timing out under load. I profiled it and found that THIR-teen database calls were happening per request — most of them redundant. I introduced a caching layer, and we got that down to THIR-teen calls total across the whole session. Response time dropped from THIR-ty milliseconds to THIR-teen milliseconds at p99, and error rate fell by THIR-ty percent over the next THIR-teen days." What the interviewer hears: 30 database calls → 30 calls (no improvement?), 30ms → 30ms (same?), 30% improvement over 30 days. The numbers blur into each other and the impact disappears.
✅ After — stress corrected
"We had a service that was timing out under load. I profiled it and found that thir-TEEN database calls were happening per request — most of them redundant. I introduced a caching layer, and we got that down to THREE calls total across the whole session. Response time dropped from THIR-ty milliseconds to thir-TEEN milliseconds at p99, and error rate fell by thir-TEEN percent over the next thir-TEEN days." What the interviewer hears: 13 calls → 3 calls (clear reduction), 30ms → 13ms (more than halved), 13% error reduction over 13 days. The story has a shape and the numbers support it.
Notice that the corrected version also swaps 'thirteen calls total' for 'three calls total' — because in context, after a caching layer, three is the more plausible number and avoids the -teen vs -ty risk entirely. This is the second tool available to you: when a number is ambiguous or hard to stress clearly under pressure, replace it with a rounder or unambiguous figure. 'Around thirty' is safer than 'thirty-three'. 'Just over a dozen' is clearer than 'thirteen' when you are mid-flow.
Interview tip
Write out the specific numbers from your three or four best STAR stories before the interview. Say each one aloud, with stress marked (e.g. 'thir-TEEN', 'THIR-ty'). Familiarity with your own numbers reduces the cognitive load mid-answer and keeps stress from collapsing.
The drill: how to fix this in 10 minutes
Ten focused minutes of contrast drilling is more effective than an hour of reading about stress patterns. The goal is to build a physical memory of the difference — so that under interview pressure, the right pattern fires automatically rather than requiring conscious thought.
Step-by-step drill (do this once, then repeat with your own interview sentences)
- Say each pair aloud three times in isolation, exaggerating the stress: 'THIR-ty — thir-TEEN — THIR-ty — thir-TEEN.' Feel where your voice rises. It should rise on the last syllable for -teen, and stay flat or fall for -ty.
- Now say each number inside a real sentence from your own experience. Use actual numbers from your CV: 'We reduced latency from THIR-ty milliseconds to thir-TEEN.' Say it five times until it feels automatic.
- Record yourself on your phone. Play it back. The difference between -teen and -ty should be audible — if it isn't, the stress is not strong enough. Increase the exaggeration until it is clearly distinguishable on the recording.
- Drill the five highest-risk pairs specifically: 13/30, 14/40, 15/50, 18/80, 19/90. These are the ones where the vowel quality is closest and the error is hardest to self-detect.
- Finish by saying your three best impact metrics from memory, with correct stress, at normal speaking speed. If you stumble, slow down and repeat until the stress pattern holds at pace.
Contrast pairs to drill (say each line aloud)
- 'THIR-ty engineers on the team' vs 'thir-TEEN engineers on the team'
- 'FOUR-ty milliseconds p99' vs 'four-TEEN milliseconds p99'
- 'FIF-ty percent improvement' vs 'fif-TEEN percent improvement'
- 'EIGH-ty requests per second' vs 'eigh-TEEN requests per second'
- 'NINE-ty days to ship' vs 'nine-TEEN days to ship'
- 'SIX-ty thousand users' vs 'six-TEEN thousand users'
💡 How to know the drill is working
Play your recording to someone who does not know what you were trying to say — a colleague, a family member, anyone. Ask them to write down the numbers they heard. If they write the right number every time, the stress is clear enough. If they write the wrong one even once, repeat the drill with more exaggeration on the stressed syllable.
How to check yourself before the interview
Self-assessment of pronunciation has one hard limit: you cannot hear yourself the way a listener does. Your brain fills in what it expects to hear, which means you can produce the wrong stress pattern and genuinely believe you produced the right one. This is not a failure of effort — it is a known property of how auditory self-monitoring works. The practical implication is that recording is non-negotiable, and a second listener is more reliable than any amount of self-listening.
Pre-interview checklist for number pronunciation
- ✓ List every number that appears in your top three STAR stories — latency, scale, team size, time, percentages.
- ✓ Mark the stress on paper: 'thir-TEEN', 'THIR-ty'. Having it written makes it easier to visualise under pressure.
- ✓ Record yourself delivering each story end-to-end, at interview pace. Listen back specifically to the numbers.
- ✓ If any number is genuinely hard to stress consistently, replace it with a rounder alternative: 'around thirty' instead of 'thirty-three', 'roughly fifteen' instead of 'fourteen or fifteen'.
- ✓ If the interview is remote, do a sound check in the same room, with the same microphone. Compression artifacts on video calls can reduce the clarity of the -teen/-ty distinction further — speak slightly more slowly and with slightly more stress than feels natural.
The goal is not to sound like a native speaker. The goal is to say a number once, clearly, so the interviewer writes down the right figure and moves on.
Pronunciation is one of the five dimensions that Mockly English Interviews scores in every mock interview session — alongside grammar, vocabulary, structure, and impact. If you want an engineer-coach to listen to your actual answers and flag where your numbers are landing incorrectly, you can book a $40 pay-as-you-go session without a subscription. For ongoing practice across behavioral and technical formats, start a 7-day free trial of Premium, which includes CV reviews and deep-dive language guides returned with specific, actionable feedback on exactly this kind of answer.
Free Interview Guide
Browse the free interview guide library first
No card needed. Practical, tech-interview-specific English guides — then start your free trial or book a single $40 pay-as-you-go session when you're ready for live coaching.
Browse the GuidesGet in front of an engineer-coach first
Start your 7-day free trial, then book a live session — a CV review, a mock interview, or a 1-on-1 deep dive — with a UK-based engineer who's actually run technical interviews. Prefer not to subscribe? Book a single session pay-as-you-go for $40, no trial or plan required.
Book a Session →Frequently Asked Questions
Tap a question to expand the answer.