I have three different WebStat trackers...Webalizer, AwStats, and Analog. Each of them have WAY different readings. AwStats seems low, webalizer is a bit higher, and then Analog shows my traffic as around 1-2 million hits per month.. Analog's hit count seems to be the closest to my AdSense impressions, but its still a bit high.
Web analytics "accuracy" is an elusive goal While you may see Google Adsense as your "gold standard", even it has built-in assumptions and limitations. I believe, for example, that it is dependent on Javascript -- so site visitors without it deployed (yes, they exist) will remain invisible. As will bots and spiders (they likewise don't use Javascript). Plus, you may not have Adsense deployed on all your pages. Every approach to web analytics is open to similar methodological criticism -- they all fall short on some dimension. So: better than chasing reporting "accuracy", aim for a high-quality tool that helps you identify actionable opportunities for improvement. Even if you're theoretically inaccurate with your measurement, does it really matter for most of us? A tool that helps you find poorly converting pages, drop off zones, high bounce search terms, visitor problem areas, etc. is worth 10X its cost and far outweighs the value of one with marginally more accurate reports. How does your current web analytics tool stack up? Is it useful only for reporting (10% of potential value) or does it help you make your website more effective (90% of potential value)?
I use sitemeter and extreme tracker. I feel like sitemeter doesn't count a lot of hits, and extreme tracker overreports my site's traffic. I have talked to other webmasters and they all have differing views. I think your best bet is just to stick to one or two analytic tools and not worry about its accuracy.
Best idea is to use multiple methods while also having statistics from other sites such as Google and Alexa.