How to drill Chinese numbers, clock times, and tones until they're automatic

July 15, 2026 · 6 min read

Ask an intermediate Mandarin learner what still trips them up in real conversations and you'll rarely hear "philosophical vocabulary." You'll hear: someone said a phone number, a price, or a time — and their brain simply couldn't keep up. The mechanical parts of Chinese are the ones conversation is least forgiving about, and they're also the most drillable. Here's how.

Numbers: three rules cover almost everything

Mandarin numbers are beautifully regular — 23 is literally "two-ten-three" (二十三) — but three rules separate learners who know the system from those who can use it at speed:

• The 零 rule. A skipped middle position is voiced as 零 (líng): 602 is 六百零二, and consecutive skipped zeros collapse into a single 零 — 3003 is 三千零三, not 三千零零三. Trailing zeros just drop: 300 is 三百.
• 两 vs 二. Both mean "two." Counting and math use 二; measure words and time use 两 — 两个孩子 (two children), 两点 (two o'clock), but 二十二 (twenty-two).
• 幺 for digit strings. In phone numbers and room numbers, 1 is read yāo, not yī — it survives noisy phone lines better.

Knowing these rules takes ten minutes. Hearing 四千零四十 and writing "4040" in under two seconds takes a few hundred reps — and because the numbers must be random (or you memorize the answers instead of the skill), a generated drill beats any fixed exercise sheet.

Clock times: 半, 刻, and the word-order trap

Times layer three conventions on top of numbers: 半 for half past (十点半 — no 分 after 半), 刻 for quarters (三点一刻 is 3:15, 七点三刻 is 7:45), and 零 again for single-digit minutes (十点零五分). Then Mandarin adds its parts-of-day words — 早上, 上午, 中午, 下午, 晚上 — which come before the clock time: 早上七点半, never 七点半早上. And the whole time phrase comes before the verb: 我早上八点半吃早饭 — "I at 8:30am eat breakfast."

Drill format that works: hear a random time at natural speed, type the digits. Then reverse it — see 7:45, say 七点三刻 out loud before the answer plays.

Tones: your ears lead, your mouth follows

The classic beginner mistake is treating tones as a pronunciation problem. They're first a listening problem: if you can't reliably hear the difference between mā, má, mǎ, and mà, your own production has nothing to calibrate against. Two drills, in order:

1. Tone identification. Hear a syllable, name the tone. Boring, fast, and astonishingly effective — most learners are far less accurate at this than they assume, and accuracy climbs quickly with reps.
2. Shadowing at two speeds. Play a word at slow speed, copy it exactly — pitch movement included, exaggerating the third-tone dip. Then match it at normal speed. Whole words beat isolated syllables: tones in context (你好's third-tone sandhi, for instance) are what real speech is made of.

Make the drills unavoidable

None of this is intellectually hard — it's volume work, which means the real enemy is friction. The drills you'll actually do are the ones that are one tap away, use vocabulary from your own lessons, and generate fresh material every session so they never go stale.

Read next