How It Works
Plain-language explanation of how PickSavvy makes predictions and what the numbers mean.
What Informs Our Picks
Team Strength
Every team has a rating that goes up when they win and down when they lose. Beating a good team counts for more than beating a bad one. These ratings carry over between seasons with some regression — last year's champion doesn't start from scratch.
Recent Performance
Recent games are weighted more heavily than older ones. A team on a hot streak is treated differently than one that's been struggling, even if their overall record is the same.
Matchup Context
Home field advantage, rest days, travel, weather for outdoor games, and opponent defensive strength all factor into every prediction. An NBA team on a back-to-back gets penalized. An NFL team coming off a bye week gets a boost.
Player Availability
Injury reports and depth chart changes are incorporated. A team missing its starting quarterback gets a different projection than one at full strength. Fantasy recommendations adjust for injury status automatically.
Usage and Volume
For fantasy sports, we track how involved each player is in their offense — target share, carry share, snap counts, and red zone opportunities. A player who sees 25% of their team's targets is a safer start than one who sees 10%.
Shooting Efficiency (Basketball)
For basketball, we go beyond basic points per game and look at the Four Factors: effective field goal percentage, turnover rate, offensive rebounding rate, and free throw rate. These predict outcomes better than raw scoring.
What Confidence Means
Every prediction comes with a confidence score. This is not a guarantee — it's a measure of how strongly the model favors one outcome over another.
The model sees a clear advantage. One team is significantly stronger in this matchup.
There's an edge, but the matchup has enough uncertainty that upsets are plausible.
Essentially a coin flip. Both teams are closely matched and either could win.
What Risk Means
Risk reflects how volatile or unpredictable a recommendation is — not just whether it's likely to be right.
Stable situation. The player is healthy, has consistent volume, and faces a known opponent. The prediction is based on solid data.
Some uncertainty. There may be an injury concern, a tough matchup, or limited recent data. The prediction is reasonable but has wider range.
Significant unknowns. Key injuries, extreme matchup difficulty, or volatile recent performance. The prediction could be far off in either direction.
How Results Are Tracked
We lock every prediction before the game starts and grade it after the game ends. Nothing gets edited after the fact.
All predictions are for entertainment and fantasy sports use only.
Sports outcomes are inherently unpredictable. Our models provide analysis and reasoning, not guarantees.