How do I measure how stable my game is?

Measure your crash-free rate (or crash rate) from real player data, the foundational stability metric, ideally tracked per version and weighted by affected players. Crash-free rate is the share of sessions or players that don't experience a crash (and crash rate is its inverse), and it directly captures stability, how reliably the game works for players, so it's the core metric for measuring how stable your game is. Measure it from real player data in the field rather than from your own testing, since stability is about what players actually experience (including on the varied devices and conditions you can't replicate), and your own machine shows the best case, not the real stability for players, capturing crashes from the field (since most players don't report) is what gives you the real-player crash data to measure accurately. Track the crash-free rate per version so you can compare builds (is the new release more or less stable than the last?) and catch regressions, since a blended overall number hides per-build stability and the regressions that matter. Watch the trend over time to see whether your stability is improving or degrading (telling you whether your stability efforts are working). And consider impact: weight by affected players so your measurement reflects how many players are actually affected (a crash hitting many players is a bigger stability problem than one hitting few), focusing the measurement on real player impact rather than raw crash counts. So you measure your game's stability primarily by its crash-free rate (or crash rate) from real player field data, tracked per version (to compare builds and catch regressions), trended over time (to see direction), and weighted by affected players (to focus on impact). A high, stable crash-free rate on real player data across your current versions means the game is stable; a low or declining one, or a new build with a worse rate, means it's not. This real-player, per-version, impact-weighted crash-free rate is what tells you how stable your game actually is for the people playing it, which is the meaningful measure of stability, far more than how the game runs on your own machine or the absence of reports.

What's a good crash-free rate for a game?

There's no single universal threshold, since acceptable crash-free rates vary by genre, platform, and audience expectations, but the more important practice is to measure your crash-free rate from real player data, track it per version, and continuously improve it, rather than fixating on hitting a specific number. Generally, you want a crash-free rate high enough that crashes rarely affect players, the closer to a near-perfect crash-free rate, the better, and a low crash-free rate (crashes affecting a meaningful share of players) clearly indicates a stability problem to address. But rather than anchoring on an absolute target, the most useful approach is relative and directional: measure your current crash-free rate from real player field data (so you know your actual stability), track it per version (so you can tell whether each release improves or worsens it and catch regressions), watch the trend (so you can see whether you're improving), and weight by affected players (so you focus on the crashes hitting the most people). Then work to improve it, capture crashes from the field, fix the high-impact ones (which, since crash volume is concentrated, removes most of your crashes), verify fixes, and prevent regressions, which steadily raises your crash-free rate. The goal is a high and improving crash-free rate maintained over time, not a one-time number, since stability is something you continuously measure and improve. Comparing against rough norms for your platform and genre can give context for whether your rate is good or poor, but your own measurement and trajectory matter more: a game with a crash-free rate that's high and stable (or improving) on real player data, with no regressions, is in good shape, while one with a low or declining rate, or per-version regressions, needs attention regardless of any benchmark. So rather than asking for a magic threshold, measure your crash-free rate properly (real player data, per version, trended, impact-weighted) and focus on keeping it high and improving by capturing and fixing the high-impact crashes and preventing regressions, which is what actually keeps your game stable for players. A 'good' crash-free rate is ultimately one high enough that crashes are rare for your players and that you're maintaining or improving over time, which you determine by measuring and tracking it well rather than by hitting an arbitrary universal number.

Best Practices for Measuring Game Stability

Quick answer: Measure crash rate or crash-free rate from real player data, track it per version, watch the trend over time, and weight by affected players. Good measurement tells you how stable your game actually is for players.

You can't manage stability you don't measure, and measuring it well is the basis of keeping a game stable. Here are the best practices for measuring game stability.

Measure Crash Rate or Crash-Free Rate From Real Player Data

Stability is fundamentally about how often players crash, so measure crash rate (or crash-free rate, the share without a crash) from real player data in the field. Real-data measurement tells you actual stability for players, not the best-case stability of your own machine.

Bugnet captures crashes from real players, so crash rate reflects real stability. Measuring crash rate or crash-free rate from real player data is the foundation, since stability is about how players actually experience the game, not how it runs on your machine.

Track Per Version and Watch the Trend

A single number isn't enough, so track stability per version (to compare builds and catch regressions) and watch the trend over time (to see if stability is improving or degrading). Per-version and trend views turn a number into actionable insight about your stability direction.

Bugnet tracks crash rate per version over time, so build comparisons and trends are clear. Tracking per version and watching the trend make stability measurement actionable, by showing whether each release and the overall direction help or hurt.

Weight by Affected Players

Not all crashes are equal, so weight by affected players, a crash hitting many players matters more than one hitting few, so measure impact by reach, not just raw crash count. Weighting by affected players focuses your stability measurement on what actually affects players.

Bugnet ranks issues by affected players, so impact is weighted by reach. So practice measuring game stability by measuring crash rate or crash-free rate from real player data, tracking per version and watching the trend, and weighting by affected players, telling you how stable your game actually is and whether each release improved it.

Measure crash rate or crash-free rate from real player data, track it per version, watch the trend over time, and weight by affected players. Good measurement tells you how stable your game actually is for players.