Bayesian posteriors
Each projection is a distribution conditioned on player history, opponent, park, weather, and the Vegas market. We publish the median and the spread.
Empire Sports Analytics
Use What the Pros Use
The pros are quants. So is this toolkit.
ESA is a Bayesian DFS simulator for MLB, NFL, NHL, and NCAAF. Posterior projections, correlated Monte Carlo, and a published-accuracy dashboard, visible to subscribers inside the product. Built for working professionals who treat DFS as a quantitative problem.
SPPaul SkenesPIT vs MIL#1CONFIRMED
Point estimate
23.4 DK pts
50% range 18.7 – 28.1 · 80% range 12.3 – 35.8 · DraftKings scoring
Two ways in. One account.
The analytics toolkit and a free baseball league, side by side.
DFS Projections
The toolkit.
Bayesian projections for every player, every game. OneTouch lineups ranked by simulation, the Lineup Engine, and a published accuracy dashboard. MLB, NFL, NHL, NCAAF.
Start free trial
Free to play
A league of real players from every era.
A baseball card collection you play. Draft real players from every era, then run a season against other managers' collections every week. Same account as the toolkit.
Enter the ballparkWhy this matters
Most DFS tools sell a point estimate dressed up as a projection. A thousand sims over a single guess produces a thousand variations of the same wrong number. ESA models each player as a posterior distribution, so the spread you see reflects what we actually know and what we actually don't. Variance in your analysis tracks variance in reality.
What this looks like in practice: when you stack, the optimizer respects how those players move together. When you size your portfolio, the simulator tells you how often the whole night underperforms, not just the average lineup. When the slate flips at 6:55, the system flips with it. Your edge is in handling uncertainty, not pretending it doesn't exist.
What's in the toolkit
Three capabilities, all in production. Pick the one that maps to how you build.
PREVIEWDK MLB Main · 7:05 ET
NYY@BOS
LAD@SD
HOU@TEX
OneTouch
Chalk 4-Stack NYYSELECTED
Salary$50,000
Proj174.2
Win %8.4%
Naked Stars
Salary$49,800
Proj169.1
Win %6.2%
Bring-Back HOU/TEX
Salary$49,900
Proj171.5
Win %7.1%
Featured Lineup
Chalk 4-Stack NYYPTarik SkubalDET$10,500
PCole RagansKC$7,800
CWill SmithLAD$3,400
1BFreddie FreemanLAD$4,000
2BGleyber TorresNYY$3,800
3BDJ LeMahieuNYY$3,000
SSAnthony VolpeNYY$3,400
OFAaron JudgeNYY$5,800
OFJuan SotoNYM$4,500
OFMookie BettsLAD$3,800
The optimizer and OneTouch share a screen: build manually or generate a portfolio with one click.
Correlation-aware optimizer
Build slates that respect correlation. An ILP optimizer with exposure, stacking, and correlation constraints, solved over the posterior, not over a point estimate.
Slate to portfolio
Slate to portfolio in one click. A diversified lineup set tuned to your contest type, archetype-mixed for tournament leverage.
Accuracy. Audited.
Every Monday, the prior slate's scoreboard. MAE, rank correlation, and position-level accuracy. Published, not pitched.
Live in the product
What you see when you sign in. Currently showing MLB, change sport ↑
CONFIRMEDESA-verified
Independently confirmed by the official source: the MLB statsapi for batting order, a DK or FD starter flag for NHL goalies, a curated depth confirmation for NCAAF starters. Highest tier.
In the table:
OFMookie Betts#3
PROJECTEDPlatform-projected
Consensus from a sister platform's projection, our rotation heuristic, or NCAAF's snap-derived usage ranking. Strong signal, not yet officially confirmed.
In the table:
SSTrea Turner#1
Every player carries a tier. Every tier is a defined commitment.
How the model works
Three components. No black box.
Correlated Monte Carlo
Lineups are scored against thousands of joint scenarios, so stacks behave like real stacks, not as independent draws.
Integer programming
The optimizer solves a constrained ILP (salary, position, exposure, stacking) over the posterior, not over a point estimate.
No hype. No flashy influencers. Just data, analysis, and hard work.
