"The peak oil problem is more to do with flow rates than the ultimately recoverable reserve (URR). The CO2 problem is more to do with URR than flow rates."

Another way to look at it is probabilistically. We're flying over the Pacific in the plane with a broken fuel gauge. There's a 95% chance there is enough fuel to land safely on island Aspo, a 50% chance we have enough fuel to land on the more distant island of Cera without having to shove passengers off the plane into the sea, and only a 5% chance of landing on island Engdahl with a full complement of passengers. To which island should the pilot head?

There is this tendency to think the mean expectation should always guide our decisions even when the lower likelihood outcomes do not have equal ramifications. To see why this is flawed reasoning just try the following experiment: Toss a coin four times. Let's say the number of heads is roughly proportional to how much oil is in the ground. The expected number of heads is 2 right? However the chance of having less than 2 heads (lower risk of rapid climate change but early peak oil) is 5/16. The chance of having more than 2 heads (late peak oil but higher risk of climate change) is also 5/16. The chance of the Goldilocks outcome (avoiding the worst outcomes of climate change and early peak oil) is 6/16. That is, getting exactly 2 heads when tossing a coin 4 times (6/16ths) is lower than not getting exactly 2 heads (10/16th).