"If students explain one calculation clearly, compare prediction to actual, and teach one visitor using a specific number, Week 4 worked."
By Week 4, students may be proud and tired. Slow down. Reduce unnecessary correction. Protect pauses.
Students teach others. They present to judges. The arc ends with them as the authority. Your job is to set the stage and get out of the way — the room belongs to the students this week.
During the Community Exhibition, redirect visitors to the students — not to yourself. If a visitor asks you a question, say: "Ask them — they know this system better than I do." If a student freezes, give them a beat before stepping in. A brief redirect is fine; taking over the explanation is not.
Do not take over. But do manage flow, redistribute visitors, notice frozen students, reset crowded stations, and help students restart. The Best Teacher vote goes to the student who taught most clearly, not the student whose bot won.
The prediction cards posted on Day 13 are resolved at the close of Day 15's final matches. Read each team's public prediction aloud against their actual Season 2 record. Missing the prediction does not mean the team failed. It means the model missed something. Find what it missed. Acknowledge the gap, then go straight to data.
Structure: Championship winner, Best Teacher winner, one specific math moment per team, one student reflection sentence, close. No long speeches. Best Teacher is determined by visitor vote — the facilitator does not choose it. The specificity is the award.
When teams share their equations and they don't match, students often look for who got it "right." Redirect that immediately: "Your equation is built from your match data. Their equation is built from theirs. Two different models of the same system. That's how modeling works — not one right answer, many defensible models." This is the most important framing of Week 4.
LO 4.1: slope-intercept equation from two match data points · LO 4.2: pre-match prediction logged before match begins, actual recorded after · LO 6.2: three-form representation — stat number, equation, MakeCode block
The 25-minute window is required for the math pressure to land. Four to six bidding rounds per team is the target — that takes time. Real-time formula calculation under social pressure, with permanent stat consequences at stake, is the most authentically mathematical pressure in the program. Cutting to 15 minutes means most teams make one or two bids and the moment never builds.
LO 1.3: percent change vs. percentage point difference, three consecutive correct dashboard entries · LO 4.2: pre-match prediction logged and compared each match · LO 7.2: revision entry with data-cited justification when >20% off
When prediction cards don't match actual records, say it plainly: "Your prediction was 4 wins. Your actual was 2. The gap is 2 matches." Then: "Missing the prediction does not mean the team failed. It means the model missed something. Find what it missed." Go straight to data. Don't soften the miss — students posted their predictions publicly and they know what happened. The analysis is the recovery.
The peer follow-up question — "How did you get that specific number?" — is the core assessment of the whole program. A student who can answer it on the board in real time, in front of the room, has the math. A student who can't will need one more round. Give those students a clear target for tomorrow's Exhibition: one calculation they can explain to a visitor without prompting.
LO 7.3: scouting report sections 1 and 2 — opponent stats as percentages and Endurance decay equation · LO 2.4: Endurance decay x-intercept predicted from equation · LO 6.3: error trace from bot behavior back to calculation · LO 9.1: a specific number or calculation named as the thing that changed their thinking
Redirect visitors to students: "Ask them — they know this system better than I do." Manage flow — redistribute visitors, reset crowded stations, and send quieter teams manageable volume first. Notice frozen students and give them a beat. If still stuck, redirect quietly: "Tell them how you calculated that number." Help students restart without taking over. Collect Best Teacher votes. Call time on blocks.
Don't explain the system to visitors yourself. Don't fill the silence when a student pauses to think. Don't step in between a student and a visitor unless the student is genuinely stuck — and even then, redirect to the student rather than answering the visitor's question yourself. Your credibility is now in service of theirs.
Visitors vote based on: clear explanation, specific number used, and whether the visitor could complete the challenge independently. Not charisma. Not confidence. Not funniest student. For the mathematical moment recognition: be specific. "On Day 9, you noticed the rival card cited your Endurance as 2 when it was actually 4, and you corrected it with the dashboard before responding. That's what Commissioners do." One real moment per team beats five rounds of generic praise.
LO 7.1: post-task diagnostic reflects Season 1 and 2 prediction practice · LO 8.2: Best Teacher award outcome reflects visitor-assessed teaching quality · LO 9.2: verbal reflection at awards ceremony (Hello Insight post-survey)