I try to follow Peak Oil (and wrote most of what little is there on Azimuth; unfortunately, Christmas period notwithstanding, I don't currently have enough consistent internet access to do the detailed checking to work more on the article ATM).
Firstly, it is in the interests of oil companies to be present things to their advantage. For example, in "all liquids fuels" figures things are aggregated by _liquid volume_ even though the energy per unit liquid volume of types of fuel (oil, ethanol, etc) varies a bit. (I actually don't know how the users of oil such as cars work in detail: is it actually the case that to go at a certain speed for a certain time a fixed volume that gets used by the engine regardless of energy content?)
Secondly, there's the issue of projections: the quickest illustrative example I can find is Iraq. Due to sanctions, various governments pre and post invasion, etc, etc, Iraqi production is way below both 1980s levels and what "the geology could allow". Here's a quote from [an article in the Washington Post probably originally written in 2006](http://www.washingtonpost.com/wp-dyn/content/article/2006/04/28/AR2006042801082.html)
> Those figures suggest misplaced optimism by Iraq's oil ministry, which in 2005 predicted crude production would reach 2.5 million or even 3 million barrels a day by the end of 2006.
AFAICS the latest figures for Iraq in the middle of 2012 are it had only just exceeded 3 million barrels. Now it's obviously the case that there's lots of stuff been happening in Iraq which would cause problems with projections, but there's stuff (both political and geophysical) going on everywhere all the time.
All the reports I've seen about how "peak oil has been averted" are along the lines of "we've found/figured out how to extract source X, _if things go as we predict_ then in 5 years time things will be dramatically different". There are various arguments about where these may be wrong in individual cases (eg, the amount of oil extracted from shale initially starts high then drops low much quicker than historical fields which isn't being accounted for) but for me the bottom line is:
1. It _certainly looks like_ the world is pretty much at an oil production plateau which is only not more obvious because economic conditions suppressing potential users of oil.
2. There are projections that various things will mean that there will be much more oil available in 5-10 years with the claim "so even if there is a problem now it's only transitory". These are projections in the same way that stock market projections are rather than projections of, say, solar output over the next 10 years. In the later case we know enough about the situation to make a good projection and reliably estimate its error bars. In contrast we don't yet have enough knowledge to know which stock market predictions are using the right factors with the right weightings, let alone how accurate they'll be.
Caveat: I've spent the last 6 years expecting petrol to get sufficiently expensive that using a car would be uneconomical and so have avoided buying one. While it's got more expensive, it's still less expensive than the alternatives and my life would be easier if I had got one. So that's a verdict on my skill at prediction.