The Duel Week Playbook

The week your startup spends in a public duel is a launch with the ambiguity stripped out. You know exactly when it starts, when it ends, who you're up against, and what the bar is - first to 100 votes on the pilot board, first to 10 verified joiners on the cohort board (R-05). The day numbers below follow the pilot board's seven-day clock; the cohort board's duel compresses the same beats into five. Most launch advice is about manufacturing attention. A duel week is about spending attention you already have access to, on a clock everyone can see.

This is the playbook, beat by beat. It assumes the duel is already live - if you're not there yet, start with what a public duel actually is.

Day 1 - the opening hour

The first hour matters more than any other hour of the duel, because early votes set the spread everyone else sees when they arrive. A 12–3 score reads like a contest; a 3–12 reads like a verdict.

Three moves, in order:

1. Post the trailer. The duel page renders a 90-second cut of your duel as one post - score, day, the stakes, the link. Post it wherever you already exist: your feed, your chat, your newsletter. Don't draft something from scratch; the trailer is derived from the live state, so it's honest by construction. (For the five-post serialized version: how to write a launch thread.)
2. Ten personal messages. Not "check out my duel" - "I'm in a public duel this week, it ends Sunday, a vote takes four seconds, here's why I think we win it." People who know the work, not the brand. Ten real asks beat a hundred impressions.
3. Answer one objection in public. Someone will ask "isn't this just a popularity contest?" in your replies or the duel feed. Answer them there, plainly. Every public objection you answer well is a vote you win twice - once from them, once from everyone silently reading.

What not to do on day 1: buy attention. Ads into a duel with a day-one budget tell the crowd (and your rival) that you can't earn it. The duel rules make bought votes worthless anyway (R-18) - sponsors can't buy a seat, and neither can you.

Days 2–3 - the grind

The quiet middle of the duel is where most duels are actually decided. The score moves slowly, the crowd watches how both founders behave, and the unglamorous work compounds.

• Five personal asks a day. Keep a list. If you can't name twenty-five people who might vote for you or join your cohort, that list - not the duel - is your real bottleneck. That's worth knowing on day 2 rather than day 6.
• Work the feed, not just the score. A duel page shows recent activity. Founders who answer questions in public look like founders who'll answer customers in public - which is exactly what a future claimer or joiner is checking for.
• Watch the other side's pitch, not their score. If your rival's one-liner is sharper than yours, say so to yourself and out-argue them in public with substance, not volume.

Day 4 - the honest check

Mid-duel, stop and look at the score rather than the effort. The arithmetic is simple: what you have, what you need, what days remain, and how many reachable humans that implies.

If you're behind, the trailer is your best card - it re-renders with the live score, so re-posting it on day 4 tells a different and truer story than it did on day 1: an underdog, a gap, a deadline. If you're ahead, defend and don't gloat - the trailing side's own pitch is what the trailer quotes, which is the duel reminding you the crowd can still flip.

If the arithmetic says you cannot get there, do not bend the rules - a second vote from the same person throws a 409 and one person counts once (R-07). Instead, be honest: the loss costs the back of the queue (R-08), not the end of the road, and the rivals you respected this week are the first users of what you build next.

Days 5–6 - the push

Now the deadline does the work. Go back to the maybes - the people who said "I'll look at it this week." Give them one sentence on what's at stake and when it ends: "resolution in under a day" is the strongest line a duel gives you, and it's true, and you got it for free.

Two cautions:

• Don't spam. One reminder, not three. A duel is public - everyone can see who's desperate.
• Don't sandbag your rival. Attacking the other founder publicly wins nothing (votes aren't counted against anyone) and reads terribly to the exact crowd deciding your week. The scoreboard is the argument.

Day 7 - resolution day

Morning: one last-call post with the live score and the close time. Afternoon: stop asking. The duel resolves at the bar or at the clock (R-05); a tie tips into 24 hours of sudden death where the next vote ends it (R-06) - if that happens, one message to your strongest five supporters is the whole game, because the side with one more reachable human wins the seat.

Either way, the result is recorded, forever, in public - a win puts the seat in your hands and the tenure clock starts (25 days on the pilot board, 10 on the cohort board, R-19); a loss puts you at the back of the queue with your record intact. The wall remembers graduates (R-14); the queue remembers nobody. What you do in the week after the duel outlasts the score.

The three rules of the whole week

1. Personal beats public, every day. Public posts set the stage; personal messages move the number. Ten direct asks will outperform any broadcast all week.
2. The clock is your co-founder. Deadlines convert maybes. Every beat in this playbook is really the clock, deployed: opening hour, honest check, push, last call, sudden death.
3. The duel is a public record, not a campaign. Nothing you do in the feed disappears when the duel resolves. Argue like someone whose next customer is reading.

The tools that go with this playbook

The board ships two free tools built for exactly this week: the Duel Day Coordinator - enter your duel's start date and get the whole seven-day beat plan as one copiable sheet, overtime included - and the Prompt Armory on the same page, five copy-paste AI prompts tuned to the board's mechanics (pitch, category, duel, claim, cohort).

If the duel is still ahead of you: readiness first, then pick your one category carefully - the machine races exactly one (R-01) - and remember why the crowd is worth winning in the first place: the first customers are hiding in plain sight.