I thought that might be your answer.
Each keyword tool has its own idiosyncracies. Generally, most tools are useful for indicating relative frequencies of searches. None, IMO, are useful for providing accurate predictions of traffic.
Overture stats are not necessarily a good predictor of the click-throughs that your serp listing will produce. Overture numbers are often skewed by the way Yahoo/Overture lumps together variations of a search, and by webmasters checking for position.
On Overture, the numbers on the most commercially desirable phrases tend to get skewed upwards exponentially, as those searches are checked the most frequently. If a phrase is three times as popular as the next phrase down, eg, it's also three times as likely to be checked, which might dramatically inflate the numbers that you see.
In any phrase area, you have to watch Overture data carefully for patterns that indicate skewing. Long and unlikely multi-word searches with high numbers, or consistent numbers from month to month, are one such indication that the numbers in that area are not to be trusted.
Additionally, these numbers are the number of searches... not the traffic that you should expect.
Where you rank, how enticing your title and snippets are, and how attractive the competition above you is, will all affect the click-through rate.
Even on Google, which has much more search traffic than Yahoo, the click-through rate at the bottom of page one might be a very small percentage of the numbers you're seeing in Overture.
[edited by: Robert_Charlton at 5:54 am (utc) on June 9, 2007]