How Does Google Detect Cloaking?

Googlebot cloaking.

[motiv8x]

I've recently heard that Google may run a few bots that crawl the web, disobeying robots.txt in order to detect sites which are "cloaking" - ie serving different content to the spider than they do to the user.

Anyone confirm or deny this?

I'm wondering how to tell if a robot is a Google-owned spoof bot operated by them to detect cloaking.

[ThomasB]

motiv8x, that's a very interresting question that probably only Google can answer. They certainly have the following variables they can play with:
- User-Agent (change it to Mozilla for example)
- IP-Address (use a dynamic DSL IP address)
- Country (spider from one of their abroad data centers or offices)
- Supported technologies (support JS, Flash, ...)
- ...

There are so many ways Google could spoof their identity that it's pretty much impossible to find out wether it's them or not. There's also a lot of good information in the Cloaking Forum [webmasterworld.com]

[oddsod]

>>They certainly have the following variables...

Urchin