FT, I think you may be ~somewhat~ correct. I changed my sites back to apc_stat=1 and there doesn’t seem to be much of a performance hit. I was going by the recommendations from php.net:
On a production server where the script files rarely change, a significant performance boost can be achieved by disabled stats.
However, it looks like w3tc performs best with apc_stat=1. I did have trouble with plugin and core upgrades with apc_stat=1 (automated upgrades with w3tc & wordpress, and manual upgrades of other php packages such as vanilla, phpbb, joomla and slideshowpro director).
It may be worthwhile to research apc_stat a bit more and see what can be done to allow the cache (either plugin or APC itself) to recognize when important files are changed, without the overhead of polling each script every time it is called. While the performance hit for a single server running a single script package may not be overwhelming, I’m sure it could create performance issues on a setup like mine, which has 30+ virtual servers running various PHP script packages on a single physical machine.
I don’t currently have the time to do a detailed performance comparison, but in general I see CPU load is about 2x greater under normal load when apc_stat=1 vs apc_stat=0. It’s still manageable at ~1.5 but I would prefer the normal load to be less than 1… it was consistently under 1 with apc_stat=0. Remember I am running 100+ PHP script packages on a single physical server with 5 IP addresses, multiple domains, Email, PHP, CGI, Perl, MySQL and PostgreSQL. So, any single measurable performance issue has exponential consequences.