› Design Preview › Aggregate Stats Bridge
Aggregate Stats Bridge
A group-level view of response patterns shown at the start of discussion — seeds the conversation with real signal without revealing who picked what. Individual answers stay private until the reveal.
When This Appears in the Round
🎴
Answer Window
Players pick their tab and lock in a response. No one sees others' choices.
🔒
All Locked
Last player locks in. Answer window closes for everyone.
📊
Stats Bridge ← You are here
Aggregate view appears simultaneously on every screen. Discussion opens.
💬
Discussion
Stats stay visible. Host moderates. Players react and share.
🃏
Reveal
Stats disappear. Individual answers shown. Matches calculated.
Mockup — Round 2 · "Your partner wants to add someone new…"
Player view (mobile)
9:41●●●
PolyCard 📊 Stats
"Your partner wants to add someone new. You haven't met them yet. How does that land for you?"
Group response · 8 players
Which deck did the group reach for?
Emotions 75%
Behaviors 15%
Actions 10%
Emotional clusters · top tab
Protective / Cautious 62%
Open / Curious 25%
Mixed / Processing 13%
GROUP CONVERGENCE
62%
🔒 Individual answers hidden until reveal. Your answer stays private.
Host view (tablet / browser)
Round 2 — Aggregate Stats
8 players · All answers locked · Discussion open
📊 Discussion Phase
Active scenario
"Your partner wants to add someone new. You haven't met them yet. How does that land for you?"
Deck distribution — 8 players
💜
Emotions
6
players · 75% of room
⚡
Behaviors
1
player · 12.5%
🎯
Actions
1
player · 12.5%
Emotional cluster breakdown
Protective / Cautious
Anxiety, Hesitance, Overwhelm, Guardedness
Anxious Hesitant Overwhelmed
5
players · 62%
Open / Curious
Curiosity, Excitement, Expansion, Delight
Curious Excited
2
players · 25%
Mixed / Still Processing
Conflicted, Not Sure Yet, Need More Info
Conflicted
1
player · 13%
Group Convergence
How clustered are the responses? High = rich discussion territory. Low = diverse → potential match-seeding.
62%
0% — all different 40% threshold 100% — identical
Host actions
▶ Open Discussion Mic
⏩ Skip to Reveal
⏸ Pause Room
🔇 Mute All
What the Stats Bridge Shows vs Hides
🙈
No names attached
The stats show category clusters only — never "River picked Anxious." Individual answers stay sealed until the reveal phase. Players can share their own answer voluntarily in discussion, but no one's forced out.
🌱
Seeds discussion, not conclusions
"We're mostly protective here — what's that about?" is a conversation. "You picked Anxious, River" is pressure. The bridge gives the host a conversation starter without assigning it to anyone.
⏱️
Lives only during discussion
The aggregate view disappears when the host moves to reveal. You can't compare your sealed answer to the cluster while discussion is happening — that would create pressure to justify or change tone before the reveal.
🔒
Cannot be used to infer individual answers
If 7 of 8 players are in one cluster and you know everyone else's tendencies, you might infer. By design the clusters are kept broad (3 max) to prevent statistical deanonymization in small rooms.
👑
Host sees the same view as players
The host's dashboard has the same data — just more screen real estate to display it. Hosts don't get individual answers before the reveal either. Keeps the power dynamic even and the game honest.
📉
Low convergence = richer discussion
When the group is spread across clusters (low convergence), that's where the most interesting conversations happen. The host is trained to treat low convergence as a feature, not a failure — "We're all over the map, let's talk about why."
🎯
Relationship to the 40% Agreement Threshold
The convergence score on the stats bridge is NOT the same as the 40% compatibility threshold. Convergence measures category clustering — a rough group temperature for this one round. The 40% threshold is calculated at the individual answer level across the full 8-week season, comparing each pair of players cumulatively. High convergence on the bridge for one round doesn't mean two players will hit 40% by season end; low convergence doesn't rule it out.
Host Script — Opening Discussion with the Bridge
🎙️
Suggested host framing when stats appear
The stats appear on every screen at the same moment the host opens the discussion mic. The host doesn't need to interpret the data — just name it and open the floor:
"Okay, so it looks like most of us landed somewhere protective or cautious — Anxious, Hesitant, that zone. A few people went somewhere more open. Nobody's wrong here. Who wants to start — what was your gut when you read the scenario?"
Or for a low-convergence round where the group scattered across all three clusters:
"We're spread all over the place on this one — which is actually the most interesting outcome. Something in this scenario hit everyone differently. River, you want to kick us off? What did you notice in yourself?"
The host never has to say "some of you picked X and some of you picked Y" — the screen already said it. The host's job is to turn that signal into a conversation.
Design Consideration — Two Display Approaches
📊
Variant A: Cluster Bars (current design)
Recommended
Broad category buckets with percentage bars. Max 3 clusters shown regardless of room size. Broad enough to avoid deanonymization, clear enough to seed discussion.
🔵
Variant B: Emoji Reaction Grid
Phase 2 consideration
Instead of cluster bars, show anonymous reaction dots — one colored dot per player, grouped loosely by emotional category. More visceral, more personal-feeling without names attached.

Does the cluster breakdown and convergence score feel right? Should the host see anything additional players don't?

Next: Full design spec → docs/superpowers/specs/