Installed the latest NSClient++ on a Windows 2000 box, added definitions to monitor it in Nagios, re-loaded nagios, and hey presto... oh wait, lots of red bits. Why? Well, Nagios couldn't connect using the check_nt command. It works fine for all the other Windows servers.
Turns out the latest version uses port 12489 by default, which isn't what check_nt (on our version of Nagios: Version 2.12) is expecting. So followed a longish process to find out what port Nagios was expecting.
The answer is: 1248
So I edited nsc.ini, set port to 1248, re-started the service and suddenly, we're all happy.
Another little amble down the road of "huh?"