If you simply calculate the average error, you could A/B test the effect of a version with reduced difficulty (e.g., with areas with have >5-10% success) vs ran

If you simply calculate the average error, you could A/B test the effect of a version with reduced difficulty (e.g., with areas with have >5-10% success) vs random. I imagine people will play longer when they get some correct. What difficulty level do you think is appealing? I was at about 5-10%. I think 50% would be more engaging. By improving difficulty, you’d have the side effect of focusing more on more important locations. Again, it’s super.