Forecast
Cleveland Cavaliers: 52.2 projected wins in 2026-27
Simulation range: between 47 and 58 (measured coverage 51.4%, not an 80%) · roster of 12 players
52.2
wins
The question
How many games will Cleveland Cavaliers win in 2026-27?
This answer is not published
- One of the data sources it depends on is not up to date, and this speaks in the present tense.
The evidence
Cleveland Cavaliers: 52.2 projected wins in 2026-27
The claim goes unshown because something specific is missing:
- · The data is older than it itself allows for speaking in the present.el dato llega hasta 2026-09-20 (3 días) y su propia vigencia es de 2
What has changed since the first version
The claim goes unshown because something specific is missing:
- · The data is older than it itself allows for speaking in the present.el dato llega hasta 2026-09-20 (3 días) y su propia vigencia es de 2
The served range is the simulation's and covers less than an 80% promises: measured coverage 51.4%, no valid evaluation
Uncertainty
The model does not give a bare number, it gives an interval. The range in this snapshot is the simulation's (10th and 90th percentiles of 800 simulated seasons with fixed strength). Measured coverage: 51.4% (n=210): it is NOT an 80%. The evaluated interval, from previous seasons' errors, is published with the next re-issue.
worse
47
wins · simulation range
better
58
wins · simulation range
What moves this number
The players contributing the most projected impact to this roster. Impact is on its real scale, not a 0-99 score.
- Donovan Mitchell+4.3 projected impact
2341 min · 2025-26
- James Harden+3.6 projected impact
2443 min · 2025-26
- Jarrett Allen+3.2 projected impact
1523 min · 2025-26
- Evan Mobley+3.0 projected impact
2079 min · 2025-26
- Craig Porter Jr.-0.1 projected impact
1147 min · 2025-26
- Peyton Watson-0.1 projected impact
1596 min · 2025-26
- Sam Merrill-0.4 projected impact
1377 min · 2025-26
- Thomas Bryant-0.5 projected impact
734 min · 2025-26
What has changed since the first version
Every roster change triggers a recompute and the previous version is kept. This is the real history, not a summary.
- 2026-08-1450.2 wins
- 2026-08-2152.3 wins
+2.16 · trigger not stated
- 2026-08-2252.3 wins
0.00 · roster change
- 2026-08-2447.6 wins
-4.73 · roster change
- 2026-08-2547.6 wins
0.00 · roster change
- 2026-08-2647.6 wins
+0.04 · roster change
- 2026-08-2747.4 wins
-0.19 · roster change
- 2026-08-2848.4 wins
+0.95 · roster change
- 2026-08-2948.5 wins
+0.10 · roster change
- 2026-08-3048.5 wins
-0.01 · roster change
- 2026-08-3147.5 wins
-1.03 · roster change
- 2026-09-0152.5 wins
+5.00 · roster change
- 2026-09-0252.5 wins
0.00 · roster change
- 2026-09-0352.5 wins
0.00 · roster change
- 2026-09-0452.5 wins
+0.01 · roster change
- 2026-09-0652.5 wins
0.00 · roster change
- 2026-09-0852.5 wins
0.00 · roster change
- 2026-09-0952.5 wins
0.00 · roster change
- 2026-09-1052.5 wins
0.00 · roster change
- 2026-09-1152.5 wins
0.00 · roster change
- 2026-09-1252.4 wins
-0.02 · roster change
- 2026-09-1452.2 wins
-0.22 · roster change
- 2026-09-1552.2 wins
0.00 · roster change
- 2026-09-1652.2 wins
+0.01 · roster change
- 2026-09-1752.2 wins
+0.01 · roster change
- 2026-09-1852.2 wins
0.00 · roster change
- 2026-09-1952.2 wins
0.00 · roster change
- 2026-09-2052.2 wins
0.00 · roster change
The real outcome
It does not exist yet: the 2026-27 season has not started. When it ends, real wins will sit here next to the projected ones, with the forecast left untouched.
How this model has performed
Measured on seasons already played, not on this one. Every figure with its window and its status.
- How often the real result landed inside the served range (the simulation's)coverage 51.4 %
2019–2025 · n=210 · not validated · Quantiles of the simulation with fixed strength: only game-to-game variance, no strength uncertainty. It is not an 80% range. Evaluated with the rosters actually used (information known after the fact): it does not accredit the real coverage or error of a preseason prediction.
- Coverage recorded on 12 August for the residual intervalcoverage 81.30 %
2021-2025 · n=150 · belongs to another interval: not attributed to the served range · The registered coverage belongs to another interval (residuals): it is not attributed to the simulation.
- Average error in wins from the season simulationaverage error 7.30
2019-2025 · n=210 · validated · ex-post rosters (information known after the fact)
- Average error in conference seedingaverage error 2.64
2019-2025 · n=210 · validated · ex-post rosters (information known after the fact)
- Average error in wins from the earlier, coarser methodaverage error 6.93
2005-2025 · n=555 · experimental
FactDerivedAssumption
What this does NOT prove
- Quantiles of the simulation with fixed strength: only game-to-game variance, no strength uncertainty. It is not an 80% range. Evaluated with the rosters actually used (information known after the fact): it does not accredit the real coverage or error of a preseason prediction.
- This is not a betting line or a market probability: it is expected wins with a range.
- It rests on today's roster. A later trade is not inside it until the model is rerun.
- The win split does not break the schedule down game by game.
From here
Contract sheet
Source
2026-09-23
Model input
2026-09-23
Computation
2026-09-20
This answer
2026-09-23
Where this comes from