Should I measure crash rate per session or per user?

Either works, what matters most is picking one and tracking it consistently. Crash-free sessions measures the share of play sessions without a crash; crash-free users measures the share of players who never crash. Both normalize raw crash counts into a comparable percentage. Consistency over time and per version is more important than which denominator you choose.

Why use crash-free rate instead of a raw crash count?

Because a raw count is hard to interpret, 500 crashes is bad or fine depending on how many players you have, and it changes meaninglessly as your player base grows. Crash-free rate normalizes that into a percentage you can compare over time and across versions, making it far more useful for judging stability and spotting regressions.

How Do I Measure My Game's Crash Rate?

Q: How do I measure my game's crash rate?

Capture crashes automatically from real players (player reports drastically undercount), then track your crash-free rate, the percentage of sessions or players that avoid a crash. This normalizes raw counts into a comparable percentage like 99.5% crash-free. Track it over time and per version so you can see whether your game is stable and whether each update helped or hurt.

Quick answer: Capture crashes automatically, then track your crash-free rate, the share of sessions or players that avoid a crash, over time and per version.

A raw crash count is hard to interpret, 500 crashes is meaningless without knowing how many players you have. Crash rate normalizes that into a metric you can actually reason about. Here's how to measure it and what to track so it's useful.

Capture Crashes Automatically First

You can't measure a crash rate without capturing crashes, and you can't rely on player reports, which drastically undercount. Automatic crash capture records every crash from real players, giving you the complete numerator you need for an accurate rate.

Bugnet captures crashes from the field automatically, so your crash rate reflects reality rather than the fraction of crashes players bothered to report. Accurate capture is the prerequisite, a crash rate built on incomplete data understates your real problem.

Use Crash-Free Rate, Not Raw Counts

The most useful form of the metric is crash-free rate: the percentage of sessions (or players) that don't experience a crash. Unlike a raw count, it's normalized, 99.5% crash-free sessions is instantly understandable and comparable over time regardless of how your player count changes.

Bugnet surfaces crash rates so you can see your crash-free figure. Whether you measure per session or per user matters less than picking one and tracking it consistently, the point is a single, comparable percentage you can watch and improve.

Track It Over Time and Per Version

A crash rate is most valuable as a trend and a per-version comparison. Tracking it over time shows whether you're improving, and tracking it per build shows whether each update helped or hurt, since a regression appears as a crash-rate jump on the new version.

Bugnet tracks crash rate by version, so you see both the trend and each release's impact. Measuring your crash rate isn't a one-time number, it's an ongoing metric you watch per version to know whether your game's stability is heading the right way.

Capture crashes automatically, then track your crash-free rate, the share of sessions or players avoiding a crash, over time and per version. It normalizes raw counts into a comparable metric.