The hosts. Both pods. One portfolio.
The 6 regular hosts only — Chamath, Sacks, Friedberg, Calacanis, Brad Gerstner, Bill Gurley. Guests excluded. Cathie Wood, Druckenmiller, Reid Hoffman and other one-off appearances don't count toward this cohort — this is “follow the people who show up every week.” Deduped to one position per (speaker, ticker, week) so Brad's NVDA on All-In and Brad's NVDA on BG2 the same week count once. Different speakers' calls on the same ticker stay separate. Shared bucket = $600K (6 hosts × $100K).
Data-driven default for aggregates. 18mo hold, -10% stop, gain-harvest +50/+100, time-decay last 9mo.
Methodology shape
Data-optimal default: moderate stop, medium hold, stepped gain harvest.
The cohort numbers above are the result of running the engine with this preset's exact parameters (stop loss, hold window, gain harvest schedule, conviction floor, portfolio mode). The recipe itself isn't published. Subscribers can see the full parameter readout and replicate the methodology end-to-end.
This is the “follow every regular host” portfolio. 6 speakers, 200+ deduped INITIATE-tier calls over 6 years, run as one $600K shared bucket. Sits in between the 4-bestie (+143pp) and 5-bestie (+149pp) aggregates because Bill Gurley's contribution is small (6 graded calls, mostly bear-OBSERVE without short language) and Brad's BG2 calls add lift but also rotation. The story: BALANCED methodology wins again — the same conviction-floor + tight stop + gain-harvest + time-decay glide pattern that dominates the 4-bestie view.