Stance drift — analysis card studies
18 payload days · 300 subjects seen · 42 carry a series · 4 crowded rollovers
source viska.signal-graph/1.0 — conviction and breadth copied from the node, never recomputed
Style A — rollover first (the finding leads)
Oil CyclePower Generation
+0.08now
12houses
↘ -0.37trend
contested3 thin of 11
Hyperscaler CAPEX ROICAI
+0.20now
10houses
↘ -0.53trend
contested1 thin of 10
Grid BuildoutEnergy Infrastructure
+0.25now
4houses
↘ -0.26trend
contested3 thin of 9
GPU Ramp
+0.20now
5houses
↘ -0.62trend
contested2 thin of 6
The CIO's corner: broad and bullish, now moving against itself. Dot radius is breadth — hollow dots are days below four houses. The line breaks where nobody filed rather than reading flat.
Style B — the standing board
AI CAPEX BuildoutAI
+0.45now
11houses
↘ -0.18trend
contested1 thin of 16endpoints disagree
FED Rate Policy
+0.12now
16houses
↗ +0.19trend
contested4 thin of 11endpoints disagree
Gold Precious MetalsMinerals
+0.75now
8houses
↗ +0.25trend
contested3 thin of 11
Strait of Hormuz Blockade RiskPower Generation
+0.50now
4houses
→ -0.01trend
endpoints disagree
Semiconductor Degrossing
+0.50now
2houses
↘ -0.15trend
1 thin of 9
Gas LNGEnergy Infrastructure
+0.33now
3houses
↘ -0.15trend
contested3 thin of 9
Copper DeficitMinerals
-0.25now
4houses
↘ -0.26trend
contested6 thin of 8
China AI Model CompetitionAI
+1.00now
4houses
↗ +0.31trend
6 thin of 8
Same row, no emphasis. Sorted by days present, so the persistent subjects sit at the top and the one-day noise never reaches the plate.
Declared gaps. No horizon axis — nodes carry major, not horizon, so a crowded bull at spot rolling bearish at 3m is invisible here. 2 of 18 payload days carry zero nodes; the series skip them.
Contract question for RES: mean is not bounded to ±1 — the corpus range is −1.25 to +1.29, so any ±1 gauge clips.