Policy debate speed reading for beginners
Speed only helps if judges can flow you. Clean at 175 WPM beats unflowable at 350. Drill articulation first, breath second, speed third.
- Speed is a means, not an end. Judges who can't flow you flow you against.
- Tags and cites get read at conversational pace. Card bodies get speed. Cross-ex and rebuttals stay clear.
- Drill articulation before speed: tongue twisters, word lists, consonant clusters.
- Build up gradually. 30 WPM increments over weeks, not overnight.
Why beginners go too fast too fast
Walk into any high school Policy round and the spreading sounds like a magic trick. New debaters assume speed itself is what wins, so they try to match the pace before they've built the articulation to handle it. The result is unflowable garbage that loses on the same flow speed it was trying to weaponize.
The actual hierarchy: clarity, then warrants, then speed. A debater reading 175 words per minute clean with tagged warrants beats one reading 350 unflowable WPM, every time, in front of any judge who isn't already a tech-debater themselves. The reason is simple: if the judge can't write it down, it didn't happen.
The structure of a Policy card and how to pace it
A Policy card has four parts: a tag (the claim, in your own words), a cite (author, date, qualifications), a body (the evidence text), and an underlined/highlighted portion (the actual functional warrant). Each gets a different pace.
Tag: conversational. About 150 WPM. The tag is the argument; if the judge doesn't catch it, the card is wasted. Make it short and crisp.
Cite: slow. 120 WPM. Author and year only; qualifications can be faster. Adjudicators flow citations as proof of evidence quality.
Card body: fast. 250-300 WPM if you're solid. The judge isn't flowing every word; they're flowing the underlined portion.
Underlined portion: medium-fast. 200-220 WPM. Slow down slightly so the warrant lands.
Tagline summary after the card: back to 150 WPM. "So this evidence proves X, which means Y for our case." This is what the judge actually writes on the flow.
Drills that actually build speed
Tongue twisters at full volume. "Red lorry yellow lorry" / "the sixth sick sheikh's sixth sheep's sick" / "she sells sea shells". Five minutes a day. The goal is clean articulation under load, not speed.
Word list reading. Take any random word list (most debate teams have one); read down it at increasing speed. Stop when you start slurring. Drill at the speed just below the slur point.
Breath control. Run a mile, then immediately try to read a card. The breath patterns you build under aerobic load translate to speech endurance.
Card-specific drills. Pick a card you'll actually use in a round. Read it three times at three speeds: conversational, fast, max. Notice which words you trip on at max. Drill those.
Recording playback. Record yourself spreading. Listen back. If you can't understand your own delivery, neither can the judge.
When to slow down
Cross-examination. Speed in cross-ex makes you look defensive. CX is a conversation; have it at conversation pace.
Rebuttals. The 2NR and 2AR are about weighing, not card-dumping. Speed those at 150 WPM. Judges write the ballot from the rebuttal, so make it flowable.
Theory arguments. T-violations and condo-bad arguments depend on the judge catching every line. Slow them down.
When the judge stops flowing. Watch for it. If their pen stops moving, you're spreading past them. Drop pace until the pen comes back.
The first three months of speed work
Month 1: articulation only. No timer. Five tongue-twister minutes daily, then read one card at deliberately slow pace, full enunciation. The goal is to make every consonant cluster crisp.
Month 2: gradual speed buildup. Add 30 WPM each week. Start at 150, end at 230. If you start slurring, drop back 10 WPM for a week, then climb again.
Month 3: integration. Practice spreading in mock rounds against a partner who's flowing. After each speech, have them tell you which arguments they couldn't flow. Those are your weak spots; drill them.
Sample lines
Want to try this against an AI that knows the format?
Practice a Policy round →