"why it is, if these two groups are so similar in the details of their data, that one of them sees nothing but Rosy Scenario while the other sees Armageddon"

1.  Estimates of depletion rates.

2.  The two groups are not so different over the next two-four years.  That's about the lead time for all but the largest or most uncertain oil* projects (e.g., pipelines, major offshore facilities, severe political instabilities, far-frontiers, etc.).  So each group can look at the announced projects and be sure that they've got all production that will come on line simply because unless something is being done in secret (Saudi?), if it hasn't been announced yet it won't be finished within the next few years.

3.  More than three-to-five years out (2008-2010) the assumptions diverge sharply: depletion rates, mythical (?) Mid-East reserves, Saudi production ability, price levels, new discoveries, and levels of new production proposed to meet what everyone agrees is rising desire for oil.

CERA implicitly assumes that there is LOTS of oil that can be profitably developed in the $30-40** range.  Just as prices stayed largely flat at ~$19-22 for over 15 years (1986-2002), they can hover around $30-$40 for at least a decade (with inflation and dollar depreciation, that's $33-44 in 2005, perhaps as high as $50 by 2010).  

By contrast, most Imminent Peakers necessarily assume that a large percentage of the $40 oil was developed between 1975 and 1985 when inflation adjusted oil prices were above $40, and that although prices were VERY low in 1998-1999 (supposedly making oil executives VERY cautious about initiating lots of new production projects), prices have been above $30 and rising for two years already so the $30-$40 oil should be sufficiently far along the development process that the projects are announced.  That is, the new production listed by HO is the $30-$40 production - and so we can already tell that there simply isn't enough of it!

       4.  CERA also assumes that just because Western consumers WANT to buy $40 oil for $40, the National Oil Companies (NOCs) that control the reservers will WANT to rapidly expand production and sell it at $40.  Far more likely is that OPEC and the non-OPEC NOCs (especially Russia) will be content to let prices drift higher.  At the very least, like Indonesia they will find it more politically expedient to ensure that their own citizens pay below-market rates for petrol than to raise domestic prices (by cutting fuel subsidies) and collect additional revenue from international sales.  

The result will be that even if total global production continues to rise gradually, the amount of oil available to Western consumers may begin to decline well before the peak.  After all, one assumption we all make is that cheap oil fuels GDP growth.  If you want to develop your country, don't sell the oil; use it to fuel economic development just like the West did.  Indonesia is the most salient example of this, but similar circumstances may occur in other large-population producer countries like Venezuela, Nigeria, Iran, and Russia.

-- E.v.T.  (aka Silent E)

Gas develops faster and so marginal gas production is more responsive to prices.  Estimates for gas production are therefore more uncertain, but the peak date is much farther off.

* All prices in inflation-adjusted 2000 dollars.  Prices are year-end crude prices from EIA.