What's the difference between crash-free rate and crash count?

Crash count is the raw number of crashes (e.g. 500 this week), a simple tally that's hard to interpret and rises as your player base grows. Crash-free rate is the percentage of sessions or players that avoid a crash (e.g. 99.5%), normalized against your player base so it's comparable over time. Crash-free rate is the better top-line stability metric; raw counts are more useful for prioritizing specific issues.

Should I track crash count or crash-free rate?

Track crash-free rate as your top-line stability metric, it's normalized and comparable over time, so it actually tells you whether your game is stable and improving, which a raw count can't (since counts rise mechanically as your player base grows). Use crash counts per issue for prioritization: the issue with the most occurrences is your most common crash to fix first. So both are useful, at different levels, overall health versus per-issue priority.

Why is raw crash count a poor stability metric?

Because it doesn't account for how many players you have. 500 crashes is alarming for a tiny game and trivial for a huge one, and the count rises as your player base grows even if your game is getting more stable, so it can mislead you into thinking stability is worsening when it's improving. Crash-free rate normalizes against your player base, making it interpretable and comparable, which is why it's the better top-line stability gauge.

Crash-Free Rate vs Crash Count

Quick answer: Crash count is the raw number of crashes; crash-free rate is the percentage of sessions or players that avoid a crash. Crash-free rate is normalized and comparable, while a raw count is hard to interpret.

Crash count and crash-free rate are two ways to measure crashes, and they serve different purposes. Using the wrong one leads to misleading conclusions. Here's the difference and when to use each.

What Crash Count Measures

Crash count is the raw number of crashes, say, 500 crashes this week. It's a simple tally. Its weakness as a top-line metric is that it's hard to interpret: 500 crashes is alarming for a tiny game and trivial for a huge one, and it rises mechanically as your player base grows, even if your game is getting more stable.

So raw crash count is misleading as a measure of overall stability, because it doesn't account for how many players you have. Bugnet captures crash counts per issue, which is useful for prioritization, but the raw total is a poor stability gauge on its own.

What Crash-Free Rate Measures

Crash-free rate is the percentage of sessions (or players) that don't experience a crash, say, 99.5% crash-free sessions. It normalizes crashes against your player base, so it's instantly interpretable and comparable over time regardless of how your player count changes. It's the standard top-line stability metric.

Bugnet surfaces crash rates so you can see your crash-free figure. Because it's normalized, crash-free rate tells you whether your game is stable and whether it's improving, which a raw count can't, making it the better metric for tracking overall stability.

When to Use Each

Use crash-free rate as your top-line stability metric, it's normalized, comparable, and tracks whether your game is getting better. Use crash counts per issue for prioritization, the issue with the highest occurrence count is your most common crash, which tells you what to fix first.

They're complementary at different levels: crash-free rate for overall health, per-issue counts for prioritizing fixes. Bugnet provides both. So track crash-free rate to know if your game is stable and trending right, and use crash counts per issue to decide which crashes to fix, rather than judging stability by a raw total.

Crash count is the raw tally (hard to interpret, rises with player count); crash-free rate is the normalized percentage avoiding a crash. Track crash-free rate for stability; use per-issue counts for prioritizing.