← All data
Forecast · Accuracy

Forecast Scorecard

How well does the POE 7-day outlook predict CPO price direction and magnitude? Scored automatically each day when actual prices become available.

Direction accuracy (7d)

63%
of forecasts called direction correctly

Direction accuracy (14d)

47%
rolling 14-day window

Direction accuracy (30d)

40%
rolling 30-day window

Independent forecasts

19
distinct anchors in 30d (4,135 raw points)

Skill vs no-change

-61%
worse than assuming no change

No-change MAE (30d)

$5.63
the benchmark our MAE must beat

Large-move accuracy

48%
direction called on 21 moves over 1σ

AI weight

50%
fixed default — awaiting scored history

Statistical weight

50%
drift + seasonality baseline

MAE (7d)

$9.97
average $/MT error

MAE (14d)

$8.35
average $/MT error

MAE (30d)

$9.08
average $/MT error

Daily breakdown

DateDirection hit rateMAE ($/MT)Forecasts scored
2026-07-2857%$7.457
2026-07-2771%$8.037
2026-07-2471%$21.147
2026-07-2386%$5.667
2026-07-2229%$7.587
2026-07-2129%$6.627
2026-07-2043%$6.607
2026-07-1714%$6.227
2026-07-1657%$6.517
2026-07-1514%$7.737
2026-07-1414%$10.127
2026-07-1343%$8.757
2026-07-1017%$11.166
2026-07-0920%$10.325
2026-07-080%$16.854
2026-07-070%$16.443
2026-07-0667%$5.333
2026-07-03100%$4.852
2026-07-020%$2.701

Accuracy is measured by comparing each forecast point against the actual Malaysia CPO close (MPOB daily in MYR, converted to USD). Direction hit = did the forecast correctly predict whether the price would move up or down from the anchor? MAE = mean absolute error in $/MT. Past accuracy does not guarantee future performance. This is an experimental model, not investment advice.

Track record transparent — now let's talk pricing. Our desk turns outlook into offers.

Get a quote →