Reading a PAR Sheet: How Operators Audit Slot Math Before Launch

Game Math · 2026-07-07 · 8 min read · By CROCO Games

What a PAR sheet contains, how RTP and hit frequency are derived from the symbol strip, the red flags worth catching, and what to request from any provider.

Every slot you will ever license has a birth certificate: the PAR sheet. It is the internal specification document where the studio defines the symbol strips, the weighting of every reel stop, the paytable, the feature triggers, and — as a mathematical consequence of all of the above — the game's RTP, hit frequency and volatility. Certification labs audit against it. Regulators subpoena it. And yet most operators sign content deals without ever asking what is in it. This guide explains what a PAR sheet actually contains, how to read the parts that matter commercially, and which red flags separate a disciplined math model from a tuned illusion.

What a PAR sheet is, and what it is not

PAR stands for Probability and Accounting Report (some studios say Paytable and Reels — the industry never agreed, the document is the same). It is not marketing material and it is not the certification report. The chain works like this: the studio writes the PAR sheet, the certification lab verifies that the compiled game behaves exactly as the sheet specifies, and the lab's report is what the regulator sees. The PAR sheet is the source of truth; the certification report is the notarised copy.

A complete PAR sheet for a modern video slot contains five blocks:

  1. Symbol strips per reel — the ordered list of symbols on each virtual reel, including how many times each symbol appears. A "reel" in an online slot is just this list; there is no physical wheel.
  2. Weighting — for weighted-reel designs, the probability assigned to each stop. Two games can share identical art and paytables and play completely differently because of this table alone.
  3. Paytable and line definitions — what each combination pays, across how many lines or ways.
  4. Feature logic — trigger probabilities for free spins, hold-and-win boards, jackpot tiers, multipliers; and the math inside each feature.
  5. Derived figures — theoretical RTP (often broken down: base game vs feature contribution), hit frequency, volatility index, max-win probability.

The word derived is the important one. RTP is not a dial someone sets at the end — it is the arithmetic sum of probability × payout across every possible outcome. When a studio ships "the same game" at 96%, 94% and 92% RTP variants, what actually changed is the symbol strip or weighting table, and everything downstream of it moved too.

The research finding every operator should know

The canonical academic work on PAR sheets is Harrigan's 2009 study, which obtained the real PAR documents for regulator-approved machines that looked identical but were programmed at different payback percentages, then ran Gambler's Ruin simulations — 2,000 simulated players with a $100 bankroll on an 85% version versus a 98% version. The result is one of the most useful facts in slot economics: the median player's session was statistically indistinguishable between the two versions. What changed was the upper tail — the number of players whose balance ever exceeded $1,000 rose roughly tenfold on the high-RTP version.

Read that twice, because it reframes how you think about RTP as a lobby lever. Payback percentage does not buy the average player a visibly better session; it buys more rare, long, big-balance trajectories. If you choose a higher RTP variant expecting players to "feel" the generosity, the math says most will not. If you choose it to generate more big-win stories and screenshots per thousand sessions, the math says that is exactly what you are buying. We unpack the commercial side of this in why 96-97% no longer differentiates.

How to read the three numbers that matter

RTP. Check the split between base-game and feature contribution. A 96% game that pays 60 points of its RTP through a feature that triggers once per 200 spins plays brutally in the base game — fine for a bonus-buy audience, rough for casual traffic. The headline number hides this; the PAR sheet shows it.

Hit frequency. The share of spins that return any win. On multiline games, remember that many "hits" pay less than the stake — the sheet will show you what portion of hits are below-stake returns. Two games with identical 25% hit frequency can have wildly different real generosity once you separate full wins from partial returns; we cover the trap in detail in our hit-frequency guide.

Volatility / max-win probability. Good sheets state the probability of the advertised max win. If a game markets a 10,000x cap, the difference between a 1-in-5-million and a 1-in-500-million chance is the difference between a real feature and a billboard. Ask for the number; a studio that will not share it is telling you something.

Red flags an operator can actually catch

You do not need to re-run the simulations. You need to ask for the summary figures and check coherence. Practical red flags:

Red flag Why it matters
RTP quoted, feature split refused Base-game experience is being hidden; expect complaints from casual players
Max-win probability "not available" The advertised cap may be effectively unreachable
Hit frequency without below-stake breakdown Marketing may be counting losses-disguised-as-wins as generosity
Near-miss behaviour with no math basis Visual near-miss tuning beyond RNG output is a regulatory risk in serious markets
Volatility label with no distribution data "Medium" is doing a lot of unpaid work

The near-miss point deserves one sentence more: on a properly built game, almost-wins occur exactly as often as the symbol distribution implies. Research on jackpot near-misses shows they produce measurable frustration-arousal in players, which is precisely why regulators in mature markets look for artificially inflated near-miss rates. The PAR sheet — not the front end — is where the honesty shows.

What CROCO shares, and what to ask any provider

CROCO provides operators the full math summary for every title under NDA: RTP with base/feature split, hit frequency with below-stake breakdown, volatility distribution, feature-trigger cadence, and max-win probability. Public game pages carry the headline specs — 777 Hell Streak, for example, publishes its 29.4% hit frequency openly. The certification trail (GLI, BMM, eCOGRA, iTech Labs) verifies the shipped build against the sheet.

Your due-diligence request list, verbatim:

  1. RTP per variant, with base-game vs feature contribution.
  2. Hit frequency, split into full wins and below-stake returns.
  3. Feature-trigger rate (spins per bonus, average and median).
  4. Max-win probability, not just the cap.
  5. Confirmation that near-miss frequency is a pure RNG output.
  6. Certification lab and report reference for the exact build you will deploy.

A provider who answers these in one email is a provider whose math team is confident. Silence is also an answer.

Frequently asked questions

What does PAR stand for in a PAR sheet? Probability and Accounting Report (also expanded as Paytable and Reels). It is the studio's internal math specification for a slot — symbol strips, weightings, paytable, feature logic and the derived RTP, hit frequency and volatility.

Can an operator request a PAR sheet? Usually not the raw document — studios protect strip-level detail as trade secrets. What you can and should request is the math summary: RTP split, hit frequency breakdown, feature cadence and max-win probability, typically under NDA.

Is the PAR sheet the same as the certification report? No. The PAR sheet is the source specification; the certification lab independently verifies the compiled game against it and issues the report regulators rely on.

Does a higher RTP make players feel luckier? Mostly no. Simulation research on real PAR data shows the median session is nearly identical across wide RTP gaps; higher RTP mainly increases the frequency of rare large-balance sessions.

Key takeaways