Add retention columns to your PAR sheets.
Your math is certified, fixed by market, and locked during design. Every decision about RTP levels, volatility, and feature pacing is made once and carried for years. Kairos uses predictions, not gut feel, to make those decisions.
Your problem
Six months, one shot
A game takes half a year and a lab fee. Volatility and feature pacing are locked before anyone has seen a player. One wrong call can burn the next six months.
RTP levels decided by debate
Operators push for lower levels. Lower RTP means higher house edge and shorter sessions, and no one in the room can say where the product of those two peaks.
A library you cannot compare
Dozens of games and daily operator summaries, but no way to explain why one title keeps players while its sibling does not.
What you get
Three new metric columns
Add predicted session length, predicted D2, and predicted wager per player next to RTP, hit frequency, and volatility in your existing PAR sheets.
RTP-level elasticity
For each certified level, calculate house edge × predicted session length, so level decisions become calculations you can show operators.
Library normalization
Normalize round data across your full game library into one model, so titles can be compared by their effect on player behavior rather than by payout outcomes.
Start with public priors
Predictions start from public priors. Round data from your operators makes those predictions truly yours. Design partners get these columns first and help shape the protocol.
Evidence
Exact enumeration, not sampling
The PAR Sheet simulator calculates RTP, hit frequency, and volatility index through exact enumeration, while also reporting Monte Carlo confidence intervals.
Open simulator769 simulated runs of a live control system
Tables determine cost; player populations determine retention; the two are almost uncorrelated. Retention needs a player model, which a spreadsheet does not have.
Read the studyThe Kairos paper
Explains how a stream of rounds becomes a language, and why the model learns the player, not the RNG.
Read the paperFirst step
- 1
Run your PAR sheet in the browser
Nothing is uploaded. You get exact math and simulated session profiles under an explicit player model.
- 2
Talk with us about your library
Which titles, which markets, and what data your operators can share.
- 3
Design partner onboarding
We normalize your round data, fit Kairos, and deliver retention columns in your next PAR sheet.
Common studio questions
Start with your next decision
Bring a table.Tell us your goal.
Want to reduce pre-submission iteration, evaluate a compensation budget, or validate an AI-generated design? Tell us which metric you want to optimize, and bring a PAR sheet, weight table, or draft. Together we’ll define the evaluation scope, deliverables, and next step.