Some more numbers.
[link|http://www.doi.gov/news/030312.htm|10.4 B barrels "technically recoverable", 1.4 Mbbl/day estimated production]. That would give 20.3 years at 1.4 Mbbl/day (real production rates will almost certainly fall after a peak).
The [link|http://www.eia.doe.gov/emeu/international/petroleu.html#IntlConsumption|EIA] has an XLS file that says the US consumption in June 2005 was 21.2 Mbbl/day, and [link|http://www.eia.doe.gov/emeu/international/petroleu.html#IntlConsumption|EIA] has an XLS file that says that the US production in June 2005 was 5.4 Mbbl/day.
Similarly, the [link|http://www.eia.doe.gov/emeu/international/petroleu.html#IntlConsumption|EIA] says that US Net Imports were 12.8 Mbbl/day in June 2005.
So if you accept all of the above, then ANWAR could decrease net oil imports by 1.4/12.8 = 10.9% - a not-insignificant amount. But, by the time ANWAR oil was on the market, it's probably safe to assume that US net imports would be larger than 12.8 Mbbl/day, giving ANWAR a smaller piece of the pie. But it may also be the case that pressure on oil markets will be larger by that time (due to increased imports by China and India) so that any increase in production will help reduce pressure on the market.
ANWAR clearly isn't a panacea, and won't be a big part of the solution to the oil import problem we have. It would help some, but, e.g., improving efficiency in transportation would help much more.
FWIW.
Cheers,
Scott.