THE ANALYSIS
A closer look
at the probability.
The original model prediction, and what it does not tell us.
Final result · SAMSUNG 7 — 8 KT
SAMSUNG
—KT
—최원태 / 대니엘
NAVER schedule · 07:25 KST checkedNo matching approved probability
No probability is calculated or adjusted in this preview.
01 / THE CIKADA MODEL
Read the evidence behind the probability.
The original approved pregame probability, alongside the records behind it.
Input cutoff · — · earlier final records
The exact input explanation for this forecast is pending. No other fixture’s inputs or estimated contributions are substituted.
TierB8 uses season game counts and win fractions for both teams, home performance at home and runs allowed over five games, away wins over ten games, and three meetings in the same home/away arrangement.
How these records are countedWin fraction is wins divided by completed games, including ties; it can differ from the standings rate. Head-to-head includes only the same home/away arrangement.
Not included in this forecastStarters, lineups, injuries, weather and expert picks are not inputs. The postseason outlook transfers regular-season form; it is not a model of current series performance.
02 / PROBABILITY, VERIFIED
Beyond picking winners.
How good were the probabilities?
CIKADA issues probabilities and keeps their errors. The Brier score averages the squared difference between forecast probability and outcome, penalizing confidently wrong predictions more heavily.
This sample is too small for a firm performance conclusion. It is neither a guarantee for this fixture nor a postseason evaluation.
Brier = mean[(probability − outcome)²]Win=1, loss=0, scored once per binary game. A 60% hit rate and a 0.60 Brier score are different quantities.
If this team wins 0.1600 / if it loses 0.3600
This illustrates scoring, not a future result. Always predicting 50% gives a Brier score of 0.25.Per-game errors behind the current score
Historical evaluations, the current daily model and expert hit rates are separate records. Expert selections below are comparisons, not weights in the model.
03 / SPOTV · SPOTIME BASEBALL
Expert selections.
See the record.
2026-03-29 — 2026-10-29
80 games · 6 commentators · 480 picks
All captured picks 294/480 · 61.3%
2026-11-03 · No verified domestic picks for this fixture yet. Records below are historical.
Partial captured-game record: repost-card samples through Sep 8 plus subsequently verified official Instagram cards. Missing dates are excluded; picks recovered after a game are retrospective records.
Original-model performance and expert accuracy are separate. Without a validated, approved weighted forecast for this fixture, the original probability is preserved.
04 / THE RECORD
Keep the evidence.
Separate the evaluations.
This view reads original predictions from a public snapshot. It creates neither expert adjustments nor backdated forecasts.
Inspect the evidence for this fixture
- Fixture
- 2026-11-03|SAMSUNG|KT|1
- Model
- —
- Generated
- —
- History cutoff
- —