Skip to content
Jacob Bets

Explore

Home

Research

DFS ProjectionsBenchmark ProjectionsIndividual StatsGame Projections

Review

Projection ResultsFantasy ChangesIndividual Line Changes

Learn

AnalysisMethodology
Contact
Jacob Bets
Learn/Methodology
Menu

Explore

Home

Research

DFS ProjectionsBenchmark ProjectionsIndividual StatsGame Projections

Review

Projection ResultsFantasy ChangesIndividual Line Changes

Learn

AnalysisMethodology
Contact

Methodology

ModelsTriangulationBenchmark

Projection stack

How each model reads opportunity

ModelApproachWhere it helps

Game environment first

Top Down Model

Starts with the expected team environment: pace, scoring, rebounds, assists, turnovers, and other team-level outcomes. That production is divided across the projected rotation based on minutes, role, usage, and historical shares.

Best for keeping player projections tied to the game script, team total, and overall opportunity pool.

Player rates first

Bottom Up Model

Starts with each player. Expected per-minute production is estimated across points, rebounds, assists, steals, blocks, turnovers, and fantasy scoring. Those rates are paired with projected minutes to create a full stat line.

Best for capturing recent form, role changes, player usage, efficiency, and expected playing time.

Reconcile the difference

Stat Gap Model

Adds the player-level projections back up to the team level, compares them with the projected team totals, then allocates the gap to the players most likely to gain or lose that production.

Best for catching projections that look fine individually but break when the full team box score is added together.

Triangulation

The edge is in the gaps

The Stat Gap Model connects the team-first and player-first approaches. If player projections add up to fewer rebounds than the team is expected to record, the missing rebounds are distributed across the rotation. If player scoring adds up above the team total, scoring is pulled back.

That reconciliation helps prevent a projection board where every player looks reasonable on his or her own, but the full team box score does not make sense.

  1. 1Build the team environment so the total opportunity pool is realistic.
  2. 2Build player-level stat lines from rates, minutes, role, and usage.
  3. 3Compare the summed player projections against the team-level forecast.
  4. 4Adjust the players most likely to absorb the gap based on historical shares and current responsibilities.
  5. 5Flag where the models agree, disagree, or create a betting and fantasy decision point.

Benchmark projections

Benchmark Projections Explained

The Jacob Bets Benchmark is a free projection set constructed to be the standard for DFS projections.

Think of it like an S&P 500 index fund. An investor can just buy the index, or use it to judge whether another investment strategy is actually adding value. A lot of active investment managers do not beat their benchmark net of fees.

The benchmark serves the same purpose for DFS projections. You can use it directly or use it to compare your preferred projections to it.

The important question it answers is this…

If you are paying for projections, or spending time building your own, what are you getting that the benchmark does not already provide?

The benchmark includes the foundational principles needed to craft quality DFS projections.

You can use it to build DraftKings, FanDuel, or Underdog Battle Royale lineups. It is free, updated regularly, and designed to cover the basics you need from a projection set.

Past benchmark projections are saved beside the actual fantasy-point results. You can download the historical files and compare the benchmark, your preferred provider, or your own model across the same players and slates.

You should not judge a projection source off one slate, good or bad. The real question is whether your chosen projections are consistently outperforming the benchmark enough to warrant the subscription cost or the time you spend building it yourself.

The benchmark is designed to be a credible baseline. It is not a guarantee, and it is not a complete DFS strategy by itself.

It does not necessarily tell you the best lineup to play, account for every contest-specific decision, or guarantee profit. Ownership, correlation, lineup construction, contest selection, and variance still matter.

A more expensive or more complex projection model may be better. But the extra cost or effort should show up in the results over time.

The benchmark gives you a starting point and a standard other projections should have to beat.

Why three models?

Agreement creates confidence. Disagreement creates questions.

Each model has strengths and blind spots. Top Down Model keeps projections anchored to the game and team environment. Bottom Up Model keeps the read close to each player's production profile. Stat Gap brings the two together so the final numbers respect both the individual role and the team-level forecast.

When all three models point in the same direction, the projection is cleaner. When they separate, the gap becomes the research target: minutes, usage, matchup, market expectation, injury impact, or game environment.

Projection Risk

Projections are estimates, not guarantees. Actual results can move quickly because of:

InjuriesFoul troubleCoaching decisionsGame flowOvertimeShooting varianceRotation changesLate availability news

Launch updates

Get notified when new tools drop

Get notified when optimizer tools, NFL projections, NBA projections, new content, and model updates become available.

No spam. Just launch updates and important projection/tool releases.

Choose updates
Jacob Bets

© 2026 Jacob Bets. All rights reserved.

Jacob Bets is for informational and entertainment purposes only.

ProjectionsAnalysisMethodologyContactPrivacyTermsResponsible Gaming