In presenting the latest data, through November 2014, Helms explained that the state had reached “record oil production—but just barely record oil production, not really a significant increase.” Looking at the history, oil production has been pretty much the same in September, October, and November. It looks like oil production is flatlining in North Dakota—and hence, in the Bakken shale.

The bigger surprise came with the data on the pace of activity in North Dakota. The Director’s Cut gives statistics that show how fast the oil boom is going, including how many drilling permits the state issued and how many drilling rigs were active. One of the most crucial metrics, I think, is the monthly count of well “completions”: that is, how many wells have been drilled, lined with cement, and then fracked—and so are ready to start producing oil.

The latest Director’s Cut estimated the number of well completions in November, the most recent month it reported, to be just 39. Earlier this year, the rate of well completions was around 200 a month—so the Director’s Cut was saying the rate of completions had dropped off a cliff.

(Note that things usually slow down each winter, but the winter of 2013-2014 was especially severe, leading to a large drop in well completions.)

Helms pointed out that, during the fall of 2014, not only did the number of active drilling rigs drop, but the places where companies were drilling shifted toward the “core” areas where wells yield more oil. Referring to a map of where rigs were located, Helms said, “you’ll see that focus into the core area is really well underway.” And yet, as Helms pointed out, North Dakota’s oil production had remained fairly flat. As Helms put it:

As the technology has improved, and as we’ve focused drilling and completions into that core area, amazingly, in November, with only 39 new wells completed, we were able to sustain production. So that tells you a lot about the core area, the fact that it’s two or three times as productive as much of the Bakken in general.

Since it appeared that only a low number of wells in the “core area” were required to hold production steady, this was a “kind of a little silver lining, or a little ray of sunshine,” Helms said at the press conference.

Helms’ count of well completions for November—just 39—was a preliminary figure. But it seems Helms was taking that number quite seriously.

Data-wrangling

However, that number of well completions was so low, such a drastic drop from the months before, that I wondered whether it was really correct—even with the price of oil having dropped by half. Also, I was skeptical that production could remain flat, when the number of completions had dropped off so rapidly. (For more on how rapidly Bakken production would drop if all drilling stopped, see my earlier article, “The Investment Gap.”)

During the press conference with Lynn Helms, I asked whether there was a lag between when wells were deemed “completed,” and when they started producing. Helms replied:

We have a legal definition of ‘completion’ which is the permanent wellhead is in place, and the well is producing into sales tanks. So by definition, there is no lag.

That kind of lag, it seems, couldn’t be the reason for production holding steady, despite so few well completions.

So I looked for another explanation. What if the reported number of well completions was wrong? What if there were actually a lot more than 39 well completions in November?

I went on a data-wrangling missing to try to get the best possible count of North Dakota’s well completions, month by month. It turned out there were actually far more well completions in November than the Director’s Cut had reported. There has been a clear drop in activity the past few months, but not nearly as large as the recent Director’s Cut stated.

According to my count, there were just over 150 well completions in November, compared with 39 North Dakota’s DMR’s count. That is, it looks to me like nearly four times as many wells were completed as Helms had said. And my count is likely an underestimate, since I excluded all vertical wells.