Prediction accuracy

Every prediction backtested against real IRCC draws. Full transparency — no cherry-picking.

Source: IRCC

Accuracy

66%

247 of 375 within range

Avg error

23 pts

Distance from center estimate

Draws tested

375

With 3+ prior draws in category

Method

EMA + slope

Exponential moving average

How we test

For each historical draw, we run our prediction model using only data available before that draw. We predict a cutoff range (center estimate +/- spread based on recent variance). If the actual cutoff falls within that range, it counts as accurate. No hindsight, no adjustments, no cherry-picking.

Recent predictions vs. actuals

Draw	Type	Predicted	Actual	Result
#411 2026-04-15	French language	387–407	419	Off by 22
#410 2026-04-14	Canadian Experience Class	499–519	515	Within range
#409 2026-04-13	Provincial Nominee	735–801	786	Within range

Draw

Type

Predicted

Actual

Result

#411

2026-04-15

French language

387–407

419

Off by 22

#410

2026-04-14

Canadian Experience Class

499–519

515

Within range

#409

2026-04-13

Provincial Nominee

735–801

786

Within range

See predictions for the next draw

Based on the same model, applied to current data.

View predictions →