Same Signal, Different Market: How News Reads Across BTC vs the S&P 500

One ledger, two benchmarks

Every bullish or bearish call we publish is graded against a real price move in our public verification ledger. But "the market" is not one thing. Crypto stories are graded against Bitcoin (BTC); everything else — macro, global equities, rates — is graded against the S&P 500 (SPY). A single benchmark would break the scoring: Bitcoin's daily range is several times an equity index's, so one yardstick would make every crypto call look dramatic and every equity call look like noise.

The neutral zone splits the same way. A 24-hour move smaller than ±1.0% on BTC or ±0.3% on SPY is classed as "neutral" and excluded from the hit-rate denominator. Only clear UP/DOWN calls with an importance score of 60 or higher get graded at all — the full scoring rules are in our verified accuracy guide.

Which news maps to which benchmark

crypto → BTC: Bitcoin, Ethereum, altcoins, exchanges, crypto regulation — the whole category.
macro → SPY: Fed policy, CPI and jobs data, Treasury yields.
global_markets → SPY: global equities, corporates, commodities.
korea_finance → SPY: Korean financial news — a small, separate track that is also excluded from the BBI.

A fair question: why is an Ethereum story graded against BTC rather than ETH? By design. Individual coins and single stocks swing on flows that have nothing to do with the story, so we ask a stricter question: did the call's direction match the reaction of that market as a whole? Two benchmarks keep the grading simple, mechanical, and impossible to cherry-pick after the fact.

The numbers: hit rates and reaction sizes by category

As of July 2026, over the trailing 90 days at the 24-hour window (decisive calls only, neutrals excluded):

crypto: 52.4% hit rate across 4,777 decisive calls. Average absolute BTC move: 2.41% on hits vs 2.18% on misses.
macro: 53.1% across 670 calls. Average SPY move: 0.59% on hits vs 0.62% on misses.
global_markets: 54.7% across 495 calls; 0.68% vs 0.60%.
korea_finance: 67.0% across 115 calls — treat this one as anecdote, the sample is small.

Two things stand out. First, equity-linked categories out-hit crypto by a small margin. Macro cause-and-effect is comparatively linear; crypto direction gets hijacked by leverage liquidations and whale flows that no headline predicts. Second, crypto's absolute moves run roughly four times larger than equities' (2.41% vs 0.59%) — the data itself justifies the split 1.0%/0.3% neutral thresholds.

Reading "vs BTC" and "vs SPY" on a pick card

The "vs BTC" or "vs SPY" label on a pick card tells you which benchmark that call is graded against. The single most important interpretation rule: never compare percentage moves across benchmarks. A +0.38% day on SPY is a meaningful move for an index; a +0.38% move on BTC falls inside our neutral zone — effectively nothing happened. The same number means opposite things depending on the market.

Related: a crypto hit typically comes with a 2%-plus move, so being right pays big — but being wrong costs almost as much (2.18% average on misses). Equity picks move less but land slightly more often. Neither is the "better" signal; they are two tickets in different volatility weight classes, and you should read them that way.

The importance score does not predict accuracy

Correction (2026-08-06). This section used to claim that accuracy rose monotonically with the importance score — 49.6% in the 60s up to 55.5% at 80+. That claim was wrong. Holding out the period after those numbers were taken and rerunning the same calculation reverses the trend completely (22.7% for the 80+ bucket). The score was never tracking how likely a call was to be right; it tracks how alarming the story sounds, and concentrating bearish calls in a rising market lowers the hit rate by construction. The full account is in When our AI was wrong. We are not printing replacement bucket numbers here — hardcoding them is what produced this error. The public ledger is the source of truth.

Same headline energy, different tape: June and July 2026

The instructive crypto misses cluster on June 1, 2026. Importance-80 bullish calls — "Strive pushes to expand its $4.2B capital raise for Bitcoin purchases" among them — were followed by BTC dropping -5.9% within 24 hours. The stories were not wrong; they were published on the eve of the June 3 crypto sell-off, and single-name good news lost to a market-wide risk-off wave. In crypto, the tide beats the boat. In the same sell-off, the bearish call on "Ethereum crushed to $1,840" landed cleanly: BTC -7.41%, a hit.

Equity misses fail differently. On July 13–14, 2026, four separate importance-90 bearish calls on "Fed's Waller signals rate hikes" were followed by SPY rising +0.38%. Hawkish talk was already priced in; the market traded the next day's softer actual CPI print instead. In equities, scheduled data outweighs rhetoric — which is exactly why the economic calendar is worth checking before you read a macro pick.

Limitations you should know

The benchmark is a proxy. A single-stock or altcoin story is graded against an index, so a call can be "right" about its subject and still score as a miss when the index goes the other way. The grading is deliberately conservative.
Category mislabels contaminate grading. An equity story misfiled as crypto gets scored against BTC. We fix these as we find them, but keep it in mind when reading the ledger.
Signal decays with time. Our 24-hour hit rate (52.9% over the trailing 90 days) beats the 72-hour rate (49.0%). News rarely carries a market past one day — and we publish that unflattering number anyway.

Browse every graded call in the accuracy ledger, and see how story weights roll up into a single sentiment number in the BBI guide.

Figures in this guide are as of July 2026 and shift as the ledger updates. Past hit rates do not guarantee future performance. This guide is informational only, not investment advice. See our Editorial Policy & Disclaimer.