Someone finally gets it. Link here. Rig counts don't matter. How long have I been saying this? Finally, someone gets it. This will be re-posted as a stand-alone post. Perhaps the only thing that needs to be read this week. Maybe this month. Stay tuned. I will post the link later. Don't worry I won't forget. It's a great article. The writer made only one mistake of any real substance. Archived.
Mr Messler begins:
I have written on shale production in the U.S. a number of times over the past couple of years for OilPrice.
My expectation that shale production would fall sharply to ~5-6 mm BOEPD by the end of 2020, has not been borne out.
My view was that a lack of drilling/completion activity due to adverse price conditions would cause production to fall sharply.
Historical annual decline rates for shale wells can be as high as 60% in the first year. [Another meme.]
In this article I discuss some of the reasons I feel this has occurred and what it may portend for oil supplies and prices going forward.
Actual production though has stayed at levels I didn't think would be possible last year. I am on record as having thought U.S. shale production would finish 2020 between 5-6 mm BOEPD.
As of the most recent EIA Drilling Productivity Report, U.S. shale production from the seven major plays has remained in a range of 7.6-7.5 mm BOEPD. [And production would be much higher if prices justified greater production.]
Why was author wrong earlier on?
- technology: drilling / completion -- extracting more oil per unit of interval than even just a few years ago
- operator high grading of their portfolios to focus almost solely on Tier 1 acreage
And then get this: rigs don't matter
Finally, production is somewhat delinked currently from the rig count, which still is less than half what it was before the pandemic. Operators have been choosing to bring Drilled but Uncompleted Wells-DUCs, online to maintain a flat to slightly rising production level, as opposed to mobilizing a lot of rigs to make new wells. In December, for example 159 DUCs were withdrawn from inventory as noted in last month's EIA-Drilling Productivity Report.
Again, the writer must be talking mostly about the Permian.
Active rigs in the Bakken are well below "less than half what they were before the pandemic." The NDIC says the Bakken has about 15 active rigs; Baker Hughes puts the number even lower by one or two. And, yet, production has pretty much plateaued; production certainly hasn't "followed' rig count.
I do take issue with the comment about DUCs. I've seen it anecdotally, and another reader has provided monthly data to suggest, that overall, in the oil basins, DUCs have not decreased much.
In fact, the DUCs that do get reported in the Bakken have come to the end of their regulatory deadline and must be completed or plugged and abandoned. Operators in the Bakken don't seem to be completing DUCs to maintain production; DUCs being reported as completed are almost entirely at the regularly deadline and must be completed. And even then, anyone following this closely, note all the wells that are "said to have come off the DUC list, and yet they are not reporting any production; most have not even been completed.
The "operators relying on DUCs to maintain production" has become a meme that is not true in the Bakken. I think that's also true in the Permian but I don't follow it closely enough to say with confidence.
The author addresses the Tier 1 concern:
One of the questions that often comes up is what will happen when Tier I acreage is drilled up. Some estimates have been put forward that this might occur within the next decade.
Rystad has challenged those estimates showing an estimate of the longevity of Tier I shale in years at present rates of drilling.
It comes as no surprise the Delaware sub-basin of the larger Permian basin is the king of shale, and operators there will retain a low cost drilling advantage for a number years beyond other plays.
Folks may want to go back and review the original Leigh Price study.
In addition, "Tier 1" is not static. Not only has the Tier 1 footprint grown significantly in the last few years, Tier 1 acreage is "getting better."
If one wants to be amazed, or let's say "woke" in this case, see the Rystad graph at the link. The initial Rystad estimate was made in 2019; the current Rystad estimate was made in 2020.
With regard to Tier 1:
Of interest also is a recent report that challenges some of the assumptions about lower tier acreage being substantially less valuable than Tier I.
In a 2019 article carried in the Journal of Petroleum Technology, a Deloitte study was showcased that showed some geological shortcomings could be overcome by application of technology of the type we have discussed in this article.
Further it challenged the assumption that rock quality alone was the determinant in obtaining maximum production from a well. We won’t develop that concept further in this article, except to note that it plays into the larger thesis that American shale production will be a vibrant contributor to the nation’s energy security for decades to come.
Wow, wow, wow -- that's been another consistent theme on the blog: the assumption that rock quality alone was the determinate in obtaining maximum production from a well. We would begin with the microseismic array but then we would have to do some real research.
Again, one may want to re-read the original Leigh Price article.
My conclusion- two things one cannot/should not predict: oil prices and the end of the world.
ReplyDeleteAgree completely. I won't predict prices, but I will suggest which way oil will be trending based on "tea leaves."
DeleteNice article and nice discussion. Some comments. Please, don't take as critical, just how nukes engage with content. [My skipper told me about being part of a group doing an inspection for an admiral on the BG carrier. The airedales and SWO-daddies came back to the big meeting with a bunch of praise. The nuke was like: dirt here, fire extinguisher out of date, since the PM folder was on the bulkhead I looked at it and they are way behind, etc. This is just how we are trained to engage.]
ReplyDelete1. Rigs do "matter". They are just non-linear. It's like your F-15. If you go from 50% power to 100% power, you don't get twice as fast. But you do get faster. So yes, throttle "matters". It's just not linear. Caveat: I know more about how the Sturgeon responds to throttle than the Eagle. So if I get an analogy wrong, my bad. But 50% Rx power to 100% Rx power gives you less than 2X (classified) speed. Hydro/aerodynamics are nonlinear.
a. High grading: As rigs are cut, they tend to leave the more marginal sites. Thus the remaining rigs are drilling better rock. This is "in general". Of course, the individual rigs can vary based on contract length and individual operator decisions and the like. But still, industry wide, when there are less rigs, you will see them concentrated in better areas.
This was obvious in the Bakken, in 2015-2016 and in 2020, with the rigs collapsing into the center of the play. You don't see any in Divide county now, right? Being in better rock, gives more oil production released per rig. But it's not really that rigs tend to MOVE into better rock. What happens is the rigs that were in good rock remain. And the rigs in bad rocks get cut. So there is STILL A DROP from losing rigs. Just less of a drop than if you cut rigs at random.
It's like firing the bottom 50% performing salespeople. Yes, revenue will drop. But not by 50%. This is also known as the Pareto (80-20) principle.
There is also an effect of the remaining rigs tending to concentrate the better crews, better equipment, etc. (Or when adding rigs/people of getting worse equipment/newbies.) But the geology is the main form of high grading.
Of course, the opposite effect occurs when we add rigs. You get “low grading”. If I double the rigs, I'm adding a lot of rigs into worse acreage. And yes acreage matters. Divide County ain't the Sanish. So, yes equilibrium production goes up when I add rigs, but less than linearly.
(2 of x)
ReplyDeleteb. Base decline: Oil wells produce more early in life and less later in life. In general. Sure, you can have individual outliers or refracks or the like. But in general, industry-wide, you have to keep putting in new wells to overcome losses from decline. If you have zero drilling (well really completions, but eventually you need to drill to complete), you will get a decline called base decline.
This is a classic issue in oil and gas production. And if anything shale wells are higher decline than conventional wells. So no, this hasn’t stopped mattering either. Of course, this isn’t the end of the world that the peak oilers want to make it out to be. But you do have to have some baseline drilling to overcome losses. Right now, overall US production is about 11 MM bopd. I estimate, we need about 450 oil-directed rigs, along with maybe 100 gas directed rigs, to maintain 11 MM bopd. This will also stabilize natural gas production. (There is a little bit of oil from “gas rigs” and even more gas from “oil rigs”.) Right now we are at about 300 oil rigs and 100 gas rigs. So we do need to add some oil rigs.
Note that base decline, itself, is not static. New wells decline very quickly (maybe 40% in first year). Older wells, decline very slowly, only a few percent a year (but are at low levels already and a small fraction of overall production). Stripper wells barely decline at all.
What this means is that the higher fraction of production coming from new wells, the larger the overall base decline is. Conversely, if you have mostly old wells, you decline slower. So, for example, California needs less rigs per bopd to overcome base decline (mostly very old wells with low decline) as compared to New Mexico (with very large fraction of production coming from recent wells. ND is in between, but more like NM than CA.
Note that base decline for a region (or the US overall) will change based on how much recent decline/growth you’ve had. So, for instance, in DEC14, base decline for the US was close to 3MM bopd/year. (Rystad estimate—and I trust them.) But by DEC16, this had dropped almost in half. What this meant then was that much less rigs were needed just to “hold serve” in DEC16 as in DEC14. Conversely of course as we grew in 17, 18, and 19, more and more of the production was extremely recent. So that rigs to overcome base decline increased. This is the Red Queen. Nothing wrong with her. She does exist. But she shows more power when you are growing fast. And actually gets less demanding after a period of shrinkage.
Note that it’s not just the level of production (e.g. 10 MM bopd versus 12 MM bopd) that affects the base decline, but the FRACTION of recent production. So, ~9.5 MM bopd in mid 2015 was a whole different kettle of fish than ~9.5 MM bopd in mid 2017. This is because the 2017 production had much less ‘new wells” as a fraction of total production. What this also meant is that it was actually EASIER to explode in late 17 and 18, than it had been in 2014-2015. Conversely, there was a bit of an air brake in 2019 and growth was not as strong (even if rigs had stayed constant--they didn’t but drop was light).
3 of x
ReplyDeletec. Level versus rigs (base decline): Even if you have a fixed fraction of new/old production, level matters. Thus, if you had the exact same distribution of new/old but doubled the level, you’d need double the rigs. An easy thought experiment is if you added a second USA next to the existing USA. Well, it would need as many rigs to maintain production as the original US, so the total rig count would be double.
Of course that is just a thought experiment. But we can consider a long term US producing 15 MM bopd versus 10 MM bopd. And if we assume same percent old/new, then we’d need 50% more rigs.
In reality, this effect tends to be smaller than the percent old/new, especially when oil prices gyrate and production grows/shrinks significantly. It does occur of course. But it is less significant than the “percent new” issue. At least for shale. At least recently. Nevertheless, despite being less significant, it is routinely confused with the percent new issue. Perhaps because it is just easier mentally to think about. (As it is easier to think of linear effects than nonlinear ones…this after all is a linear effect, remember our duplicate USA thought experiment.)
d. Time lag: Adding rigs does not instantly change production. It takes time. So if you have speed X at 50% throttle, you don’t instantly go to speed Y at 70% throttle. There is a process called acceleration. Acceleration changes instantly with your throttle. You can feel it even. But air speed takes time. Of course, eventually, you will end up at whatever is the equilibrium speed (where thrust equals air resistance). But you don’t get there instantly. This seems crushingly obvious when driving, flying, or the like. But you still see people routinely confuse this in discussions of oil production. It just takes time, to get to the new equilibrium. You can’t judge things off of one week or month or the like. Even leaving aside all the other confounders like DUCs.
4 of x
ReplyDeletee. Equilibrium level versus increase/decrease in rigs: An increase in rigs only leads to an increase of production if you are above the level needed for base decline. Let’s say you need 450 rigs to maintain production. If we are at 300, we might add 50 rigs, but still decline. This doesn’t mean rigs are useless. It just means you need to be higher than 450 (or whatever the number is). It’s just like flying. If you are going 600 knots at throttle X, and you go to zero, you will start slowing down. Now, if you move the throttle to less than 0.5 X, you might still keep slowing down. You won’t slow down as fast. But you will keep slowing down. This doesn’t mean throttle is useless. I routinely see peak oilers (and shale boosters) messing this concept up.
f. DUCS: DUCs are a curious thing. At equilibrium, you will have some percent of DUCs at any time, given the delay between drilling and fracking. This varies from play to play and project to project, but a typical delay might be about 5 months or so for modern shale wells. Thus if rigs drop, there is still an inventory of wells that can be completed.
Drilling is about one third of the cost. Fracking is about two thirds. What this means is as prices drop, the justification for drilling a new well (spending for drilling and completion) will fall faster than the justification for completion. Remember CLR a few years ago saying “we’ll drill new wells at $60, but we’ll complete DUCs at $50”. Thus as price drops, production will fall a bit slower than expected from rig count. Because there is this inventory of completeable wells.
This is a temporary phenomenon of course. Eventually the completable excess DUC inventory gets cleaned out. (There are some “rotten DUCs” of course. Also some new DUCs from remnant drilling, but I’m talking about the inventory of excess DUCs that will get done at the new price, but wouldn’t be drilled now.)
This can be confounded by issues with contracting length (more rigs are on annual contracts, whereas completion tends to be more ad hoc purchased). But IN GENERAL, the impact of DUC inventory is to slow decline. You still reach the SAME new lower quasiequilibrium (for a lower rig count), but it just takes longer.
Note, of course, that the opposite effect can occur during a boom. If I double the rigs, it takes five months or so, to start seeing the benefits. This can be even longer if pumping equipment goes into short supply, during a boom. Again, this doesn’t change the new higher quasiequilibrium LEVEL. Just to time to get there.
5 of x
ReplyDeleteg. “Technology”: Over time the industry gets better. Shale is still a relatively new phenomenon. We find ways to do things cheaper. We try new methods (some work, some don’t, but the ones that work eventually win out). In some cases, there’s just a “practice” effect of getting better at doing the tasks, especially for younger workers.
Note this is not an instant effect. The best way to think of it is as a slow grind of improvement. Not as discovering new drugs or superconductors. But it really is happening and is important.
I would argue that the pace has definitely dropped, compared to 2010-2016. But it’s still happening and helpful. A noteworthy, relatively late, example is the Haynesville renaissance. Even the recent Bakken increases are notable (both being “older plays”…maybe in some ways benefiting from learning from the younger plays.) But don’t oversell it either. I mean we haven’t had a Barnett/Fayetteville renaissance have we?
Note also, that when you see EIA’s DPR double during a price crash, that does NOT mean technology doubled. The time frame for tech is much slower. What you are seeing in a few months is high grading, not technology.
Also, of course, when rigs get added fast, out of a crash and DPR drops, we didn’t suddenly get shitty at technology. Yeah, there are some “green hats” and worse equipment coming into the plays…but the big impact is geological low grading. That is what you are seeing.
h. “Inventory”. Oil is a non-renewable resource. As a play is drilled out, inventory of drillable locations drops. Now, yes, we do find new areas with time. The EF came in later than the Bakken. And the Permian came in late (and the Delaware within that). And the Utica (gas/cond play, not the oil we hoped for) came in late. And even within plays we learn things. But arguably new information has been relatively slow since 2013 or so. And we have/are drilling out some of the best land. There’s a reason why EOG is concentrating more in the Permian and less in the Bakken and EF. They had GREAT land in EF/Bakken. But…they drilled it. That does happen. It’s not the end of the day. I actually think if you just take “technology” and “inventory” and cancel them against each other, it’s close to right. But you can’t just ignore it. Can’t say it doesn’t occur. That there are zero head winds.
Note also that, “child wells” are not as good as “parent wells”. It’s not some peak oiler end of the world catastrophe. But it’s also not some cheerleader irrelevant factor either. And that it doesn’t really matter that much if you drill them at the same time or over time. If you increase the density, the per well production drops (regardless of timing). Now, it can still be useful…sort of like adding rigs themselves. But there’s a headwind. CLR had great slides explaining the concept. Every added well helps the overall production, but drops the per well production. Since the wells cost money, you end up with an optimization problem (like a maximum in freshman calculus). But basically child wells are not as good. You can actually think of child wells as Tier 2 hiding inside the Tier 1.
Also, yes, with technology (broadly defined to include cost savings), we can sort of consider Tier 2 to move to Tier 1. But we already took CREDIT for technology. So, to just completely ignore depletion doesn’t make sense.
I prefer my simple-minded rough technology and depletion counteract idea. It pushes back on both the overenthusiastic cheerleaders and the overgloomy peakers. Check out Trisha Curtis of Petronerds, she has similar idea, I think. Realistically what this means is good for the cornucopians. But don’t be over-cornie. Depletion is not completely non-existent.
6 of x
ReplyDeletei. Rigs versus frack crews: In the SHORT term what matters is fracking crews since they really lead to oil. But in the LONG term, every completed well needs to have been drilled. So if you care about the long term impact of production for ND and the US (and I think you do), then rigs matter. Once you work through excess DUC inventory, you need rigs. You just do. It’s math. It’s physics. The sea is a harsh mistress.
So, if you are looking at the long term issues of oil independence, watch rigs. Frack crews will come up/down, eventually. But every DUC needs to be drilled. If you want to know where production is headed to long term, look at rigs, not frack crews.
There is also an issue of data quality. Rig count has very high quality, granularity. Rig count is very poorly tracked and has a lot of modeling inside the numbers. Granted, this can be like looking for the keys under the lamppost. But I think you can understand the appeal of a metric that has high quality, low lag versus another with a lot of uncertainty/modeling. Well, I hope you can.
7 of x
ReplyDelete2. The Price study is pretty old in the tooth now. More recent studies have been like in the 200s B bop. I know you like the cornucopian things. And I’m even broadly on your side. But be a critical thinker please. I trust the more recent assessments more. Don’t just believe in the things were you like the answer. Be a thinking observer.
Note also that oil in place is pretty different than recoverable oil. Maybe 50 years from now, we recover it all (or 50% of it). But given current techs and reasonable medium term extrapolations, you need to take a huge haircut on oil in place versus recovered oil. This still leaves you a lot of oil. And if you take USGS at about 10 BBO and just double it (from being old, etc.) to 20 BBO, it’s still a metric shit ton (submariner unit of measure). But it ain’t 450 BBO either.
And it ain’t the Permian. It just isn’t. I know you want it to be. But it ain’t. Try to be a critical thinker versus a cheerleader. USGS is putting the Permian at 70+ BBO. And that is recoverable. Not “in place”. Oh…and the Persian Gulf has even more. You need to be apples to apples. Don’t compare oil in place (Bakken) to recoverable (Permian). It’s just not intellectually honest.
3. Transport: The Bakken is a great play. But you can’t teleport it to TX. You can’t teleport the CA oil sands either. It costs money to move stuff. And the Bakken faces a headwind on transport. It just does.
And don’t you dare say DAPL doesn’t matter. If that gets shut down…it will be a cost impact. And even the possibility that that happens is making people slow down on adding rigs, and yes, they fu…ricking matter (sorry, attack boats are all male and live in the 50s). If DAPL was assured and Keystone were approved 100%, you’d have 80 rigs running. You really would. And they matter.
DC is out. Not the video. Grr...new girl...grr...
ReplyDeleteI just saw that. I will post it after I take Sophia to Tutor Time.
Delete