MisleadingCharts
Back to the gallery

The 20% porosity that was the best 10 feet of 681

Showing the misleading chart

Be the first to star this exhibit

A prospect slide we drew for an invented exploration company, from a real 1996 Kansas wireline log, quotes 20% porosity for the Lansing–Kansas City in an offset well. The figure is right for 4,530–4,540 ft. It is also the best of 1,343 possible 10-ft windows in a 681-ft interval that averages 6.2%.

01The claim

An offset well logged 20% porosity in the Lansing–Kansas City, so a prospect on the same trend should find good reservoir there too. A prospect slide like this one, which we drew for an invented exploration company from a real 1996 Kansas wireline log, shows that 20% on a log zoomed to 30 feet and books the interval as a primary objective.

02The trick

Every reading on the slide is real. Between 4,530 and 4,540 ft the density porosity averages 20.0%, the gamma ray stays at 21–28 API, which means clean, low-clay rock, the neutron log agrees at about 19%, and the caliper holds steady at about 8 inches, so there is no washout to blame. What the slide leaves out is everything else. The headline names a whole interval, the Lansing–Kansas City, which in this well runs 681 ft, from 4,302 to 4,983 ft. The chart shows 30 of those feet and averages 10. This is the cherry-picked window turned on its side: on a well log the window is a stretch of depth rather than time, and the effect is the same. We slid a 10-ft frame down the whole interval in half-foot steps. That gives 1,343 possible windows, and the slide’s window is the highest of all of them (they overlap, so the windows either side score almost as high); the median window averages 5.1%. Nobody needs to be dishonest to land there. Whoever builds the slide scrolls to the most interesting part of the log, and that is where the porosity is. The zoom then makes 10 ft fill the whole frame, so the eye can’t see how small a share of the interval it is.

03The fix

Draw the whole interval the headline names, at a scale where its full thickness fits, and box the window inside it. Here the 10 ft becomes a thin red stripe, 1.5% of the column. Across all 681 ft the density porosity averages 6.2% (median 5.0%), and 580 ft read under 10%. Then say how the porous rock is spread out. The 101 ft at 10% or more come in 16 separate beds, two of them about 24 ft thick, and the slide’s window sits inside one. That bed is real and worth noting, as a bed. A gamma-ray cutoff takes the count down again: 26 of those 101 ft sit where gamma ray reads 45 API or more (a cutoff we chose), in shalier rock, where density porosity also counts water bound in clay rather than usable pore space. That leaves 75 ft at 10% or more in clean rock. Porosity also isn’t pay: it says how much room the rock has, not what fills it. That needs the well’s resistivity log, a separate file in the KGS archive, and a saturation calculation (the well itself was completed as a gas well in the Morrow, more than 1,200 ft deeper). Whenever a figure is quoted for a named interval, a period or a region, check whether it was measured over all of that interval or over a window inside it, and ask how the window was chosen. A window picked after looking at the data is closer to a maximum than to an average.