Methodology
SympathyPlay exists because the market prices the obvious winner in minutes but takes days to price the rest of the chain. Our job is to map that chain — fast, structured, and honestly scored.
1 · Classify
News flows in from wire feeds every two hours. Each item is scored 0–100 for significance: how much real capital does this event redirect across a supply chain? Fluff, opinion, and minor updates are discarded. Only material events proceed.
2 · Map the chain — all the way down
Our analysis engine (built on Anthropic's Claude) drafts the value chain under strict rules: organize exposure into tiers, include second-order plays, include who gets HURT, and always end with a 'Deep cuts' tier — 3–6 smaller or obscure public companies further down the chain (equipment, materials, testing, services) that a generalist wouldn't think of. Every company carries its full reasoning chain (event → X → Y → this company), an estimated revenue exposure (labeled as an estimate), a one-line bear case, and three specific research starters. The engine is explicitly forbidden from inventing tickers — uncertainty must be declared, not papered over.
3 · Verify tickers
Every ticker is checked against a database of ~2,000 verified major-exchange listings (sourced from SEC data plus international supply-chain names). Anything we can't verify is flagged 'unverified' visually — never silently dropped, never silently trusted.
4 · Human review
Nothing publishes automatically. A human reviews every draft — edits theses, removes weak calls, tightens the contrarian take — then approves. AI drafts, humans decide.
5 · Score in public
When a map publishes, we snapshot the price of every mapped company. At 7 and 30 days, a scoreboard appends to the map: 'Map said SMCI = strong benefit → +14% since.' Hits AND misses. The aggregate lives on the track-record page, where anyone can audit our hit rate.
What we are not
SympathyPlay is not an investment adviser and maps are not recommendations — we are a research gateway: every company card is a starting point with its reasoning laid bare so you can attack it. Exposure ratings are qualitative judgments about business relationships, not price targets; revenue-exposure figures are estimates, not disclosures. Deep cuts deserve extra skepticism: smaller companies mean lower liquidity, wider spreads, and thinner coverage, and AI-drafted reasoning — even human-reviewed — can contain errors. Markets are adversarial; being mapped does not mean a stock will move. We publish our hit rate precisely so you can calibrate how much weight our analysis deserves — see the live track record.
Questions, corrections, partnership ideas? Every map has a feedback box — or write to us at the address in the footer of any email.