← Back to the dashboard

Forecast recap — 2026-W38

In week 2026-W38 the BTC Event Oracle resolved 297 calls; 35% of directional calls were correct, coverage was 24%, and the average price error was 1325%.

Every call below was graded against a naive random walk — the “no change” benchmark. Short-horizon Bitcoin is close to a random walk, so the honest, expected result is to roughly tie it; we publish the week as it landed, misses included. See the full live record on the accuracy page.

The week in numbers

Resolved
297
Hit rate
35%
Coverage
24%
Avg error
1325%

Hit rate is the share of up/down calls that landed correctly (50% = a coin-flip). Coverage is how often the real price finished inside the stated range. Avg error is the mean % by which the central price guess missed the realised price.

Every call, graded

HorizonSaidActualEst.RealisedRangeCall
1 Monthupup$73,209$81,436out of range✓ hit
1 Weekupup$77,286$81,436out of range✓ hit
1 Monthupup$75,590$81,436in range✓ hit
1 Monthdownup$74,164$81,436in range✗ miss
1 Monthupup$74,820$81,436in range✓ hit
1 Monthupup$75,043$81,436in range✓ hit
1 Monthdownup$75,049$81,436in range✗ miss
1 Monthdownup$75,121$81,436in range✗ miss
1 Weekdownup$77,213$81,436out of range✗ miss
1 Monthdownup$76,266$81,436in range✗ miss
1 Monthupup$76,451$81,436in range✓ hit
1 Monthdownup$78,068$81,436in range✗ miss

A ✓ hit means the directional call matched what actually happened; ✗ miss means it didn't — the ✓/✗ text carries the meaning, never colour alone. At short horizons Bitcoin behaves close to a random walk, so a mixed week is the normal, honest outcome rather than a broken model.

Keep reading

← Back to the dashboard