Too late to buy?

All the different ways of solving the shelter problem. To be static or mobile? Roots, legs, or wheels?
User avatar
Sclass
Posts: 2791
Joined: Tue Jul 10, 2012 5:15 pm
Location: Orange County, CA

Re: Too late to buy?

Post by Sclass »

I should temper my commentary by saying given any amount of error bar, the supply of oil is finite and therefore there will be a peak. We are in big trouble if we maintain the current path. My complaint is the condfidence in the prediction of when. The estimate of reserves can also smooth or sharpen the peak.

To shed some light on how many assumptions are made in calculating reserve volume in a given field I can illustrate the process in lay terms for the curious. A simple oil reservoir is like a piece of layer cake that you crush from two sides to create a little turtles back in the top of the cake. You fill some of the permeable layers with salad oil and water and the fluid is held down by an impermeable layer of jelly or whip cream. Since the oil and gas rises up In the mixture it spills out at the edges of our fold and oozes up and away. So we have water, oil then gas trapped under a dome. It's trapped in porous sandstone, limestone, shale or some combination. We can estimate the volume of the reservoir based on seismic images, well log data and some knowledge of local geology from surface studies. You map the extent of the dome and the thickness based on well and seismics.

Your electrical data from the wells tells you how porous your rock is...anywhere from 30-0%. You assume it is about the same between your wells because your seismic image doesn't see porosity. Your well logs don't see permeability. You assume your oil saturation and oil water contact is continuous between your wells along the geometry you've inferred from your seismics. You hope you have a layer cake down there that has just been scrunched up a little. Multiply porosity times volume and you get oil content. Hopefully you have something simple like a layer cake down there. This is a gross simplification of reality but it's all you got.

Because of the thickness of reservoir bedding, extents of the subsurface oil water contact, and homogeneity of the rock you make a lot of guesses. It's kind of like trying to guess the percent water in an unborn child using ultrasound and a poke or two of an amnio needle. And seriously 3d ultrasound is a way better image.

Note the number assumptions. We literally trying to guess about things we cannot see or touch 1000s of feet below the surface. All measurements are indirect. We measure rock geometry by reflected sound from impedance contrasts in the earth. We measure void volumes (porosity) using electrical resistance of water saturation or radioactive emission from H atoms. Big uncertainties just in the actual measurements because the petrophyscal relationships are non unique...it's not a perfect earth. Each well is expensive so you don't have many "reality checks " to contrain your seismic images. You may have some cores but the curse of cores is they are only like a straw punched into the cake at a few points (very localized data).

So you guess. Your seismic nor well log data won't tell you permeability across your field. You can have ideal 30% porosity (which you estimated empirically from electrical resistance which is incidentally non unique ) but no pore connectivity to pump oil to the wells you drill. Not to mention the small kink of faults. Faults can form impermeable barriers in the reservoir or worse, natural fracs that bled your precious oil out of the reservoir eons ago. You will miss those with sparse exploratory wells and low res seismic. You assume you have a continuous body of oil between your wells where you see it along the rock body that reflects brightly on your seismic image. You have no way of knowing what is really down there. It's just a guess. The more plate tectonics in your area, the more jumbled up the picture gets. Take your cake, cut a few slices through it with a knife, heat it up, cool it, smush it around a few times, age it and now you don't know what is down there.

Now you have to tell your boss that? Or do you tell him how many kbbl you have in place so he can tell the accountants how much to pay or amortize this bill of goods for? Do the former and head to the bar and watch some widescreen.

After actually going through this exercise in two fields in my youth I cannot believe how anyone could put a price tag on a field. What I didn't see is the big dogs hedging and sandbagging their offers based on the predicted reserves data. I bet that is an art and science in itself but I never knew the finance department people well enough.

So that is where I'm coming from when I say it is really difficult to pinpoint peaks based on reserves that are calculated to the best guess of some oil business folks or USGS govt workers.

Earth science is messy. Hope this gives some indication of just how little is known about fluids in place. I was so frustrated with this job I quit and went to grad school in a completely different field.

jacob
Site Admin
Posts: 15907
Joined: Fri Jun 28, 2013 8:38 pm
Location: USA, Zone 5b, Koppen Dfa, Elev. 620ft, Walkscore 77
Contact:

Re: Too late to buy?

Post by jacob »

@SClass - Dude, you're doing the incredulity thing :-P I think I explained how to deal with the uncertainty of reserve estimates rather well in the link I posted. At least the professional oil people were impressed back then. So look there for detail. Here's the short story.

I'll agree that we start with an estimate of reserves along with an error bar.

That is step one. So far so good. But it's definitely hard to make any predictions with these numbers, so we need to analyze a bit deeper!

Step TWO plot the reserve estimate as a cumulative probability function. From this you can get the actual mean, called the P50 (probability=50% that reserves are larger than this number). This is not what oil companies (in the US) traditionally do. They use P90. But it is what they SHOULD do if they want an accurate number. Note that the estimate can still be wildly inaccurate.

Step THREE when new data arrives, the estimate is updated and BACKDATED. Again, this is not what oil companies traditionally do. If the estimate increases oil companies show this as a new discovery. This is highly misleading because it's not new found oil but merely correcting a previously wrong number, like if I found an extra $10 in my wallet ... that doesn't mean I just made $10 ... it means I always had it but I just didn't know it yet. Hence instead, you need to update your original estimate. Thus updating you'll see that your precision is increasing! The size of your original error bar is decreasing! And because you're focusing on p50, you'll stay accurate.

Step FOUR plot the cumulative discovery (you have the numbers from step 3) for all fields as a function of year. This is called a creaming curve (a concept invented by Shell way back, google it). You'll see that this plot converges asymptotically towards a fixed number, the total amount of reserves in the ground!

Step FIVE from this curve, subtract the cumulative amount of oil (by the year) you've pumped out. This number is rather well known. You now have the amount of reserves left in the ground.

So far we haven't left high school statistics. We've just used the available data in a way that reflects the geology of the situation rather than the profit/accounting of the situation.

Step SIX Use the insight from the central limit theorem that says that if you have a large number of individual production rate curves with idiosyncratic shapes, the sum of their total rate is normal (Gaussian). The time-integrated rate curve is obviously the total possible production. Which you know from step four. So now you have the correct scale!

Step SEVEN Identify the midway point (x-axis, the time) of that Gaussian. That's the peak year. (You can make it more accurate by grouping the fields into countries or major regions ... and then adding those than merely bunching everything together ... but you get the point.)

This method identified both the US peak and the global peak within a few years of precision. So my confidence is derived from the empirical observation that the analysis worked out to be pretty accurate. Given that all the steps made rational/logical sense (in other words, there was a "sensible narrative") I'd also say that it wasn't pure luck. IOW, I'm rather confident that if the same method was used again on e.g. coal, natural gas, gold, etc. it'd work again too. Also note that at no point was did we use any kind of math more sophisticated than what's taught in junior high except the CLT which is taught to freshmen. The only difference was in the painstaking work of (A) creating the cdf reserve figures out of the error bars; and (B) properly backdating all updates. Gathering that data is a huge piece of work, but once done, it should be relatively easy to understand steps 1-7.

User avatar
Sclass
Posts: 2791
Joined: Tue Jul 10, 2012 5:15 pm
Location: Orange County, CA

Re: Too late to buy?

Post by Sclass »

Thanks for the insights. Steps 1-7 are where the problem left my hands and my understanding of this process is weak. A cursory glance of your description makes reasonable sense. I'll read it again a couple of times.

For years I scratched my head about how the reserve data could be used to come up with a price for a particular field. The ones I worked on were bought and sold over the last hundred years multiple times by minors and majors. Undoubtedly for good and bad prices since many were forced to divest. Just because I didn't get it doesn't mean it wasn't done in a logical way. Nobody seemed to get caught holding the bag of depleted wells before the oil price got 'em.

Shell USA eventually lost its dominant role in Southern San Joaquin. But it was more of a predicting the price of oil issue rather than a calculated reserves issue. In fact, every field they owned is still being pumped in some capacity. The funny thing you mention about finding the ten in your pocket is how this could happen in an older field where some guy in research comes up with an enhanced production trick and it changes the score. Sometimes it was as easy as changing a few of your injecting wells into producers to get at previously unrecoverable oil. Now the hydraulic fracturing is doing it on surprisingly many types of rock, not just shales...but at a damning price as you said.

Incredulity thing? Yessir. I've been stung by too many "stingless" bees. :o Makes me psychotic and paranoid.

Your comments on convergence are particularly disturbing. Yes, given enough time we will undoubtedly get to the right answer. It is a slow motion train wreck.

User avatar
Sclass
Posts: 2791
Joined: Tue Jul 10, 2012 5:15 pm
Location: Orange County, CA

Re: Too late to buy?

Post by Sclass »

jacob wrote:
So far we haven't left high school statistics. We've just used the available data in a way that reflects the geology of the situation rather than the profit/accounting of the situation.
!-7.
Can you explain this in a little more detail? I'm not attacking you, I just don't understand how price can be decoupled from discovery rate. My best pal at Shell was a Paleontologist and he was the first to go when exploration activity got throttled down during our cash crunch. It's burned pretty deeply into my mind.

In his farewell party he gave a speech to the effect of, "carry on fellows you're flying blind." We primarily focused on production activity from that point on. Searching with Dr. Dino was expensive.

jacob
Site Admin
Posts: 15907
Joined: Fri Jun 28, 2013 8:38 pm
Location: USA, Zone 5b, Koppen Dfa, Elev. 620ft, Walkscore 77
Contact:

Re: Too late to buy?

Post by jacob »

Price is so much harder to predict because it's [by definition] the market clearing intersection between the supply and the demand curve. Geology puts a limit on the supply, so we can predict that curve fairly well. This is the whole argument above. However, demand depends on humans and they economy. In the pre-peak days, people were very much of the 1970s orientation that demand was stiff (inflexible) and thus any decrease in supply would cause a rapid increase in prices (like the run up in 2007). However, it turned out that demand was rather soft due to the enormous amount of waste in the system. Hence people can easily reduce demand when faced with high prices and that is enough. Pre-peak we could play with increasing supply as well, but that's no longer the case. The present understanding is that the system functions more like an informational system, like a phone line, in that transmission, here price, becomes very volatility once capacity is hit.

There is some work in terms of understanding the "energy cost" (EROEI, google it) of extraction compared to the "paper cost" (money), but those numbers are strong estimates since it's hard to measure exactly how much energy goes into producing a barrel. For example, do you include the food used to feed to roustabouts? How about the energy used to build the tractors ploughing the fields where the food is grown. There are of course various methodologies, but one can only make general statements. E.g. conventional oil used to be 25+ ... an investment of 1 yields 25 ... this is obviously a good number even if it's 20 or 40% off, it's still good. Conversely, shale oil is a measly 5 and would require a lot more resources to get the same yield. In particular, if that number is off, it might not work out at all. Which is apparent now. But predicted long ago. Biofuels and solar numbers are equally terrible, so it doesn't bode well for them either. This is why they're subsidized e.g. need to commandeer resources from the rest of in order to work at all.

PS: Incidentally, the IEA, the official representation of the energy industry, surrendered to the peak oil people in 2011. There was a brief moment of "I told you so" celebration.

User avatar
Sclass
Posts: 2791
Joined: Tue Jul 10, 2012 5:15 pm
Location: Orange County, CA

Re: Too late to buy?

Post by Sclass »

I still don't get it. I was hoping you were going to say something easy to understand like price of oil is in the top of the discovery equation and the bottom and it gets normalized out. Or it subtracts itself out of the discovery rate equation.

My naive keyhole view of the industry told me as the price went down we stopped searching. When we stopped shooting high res seismics, our knowledge of the geology was restricted to that which was already done with archaic instruments. So the geology of the earth hadn't changed, but our chances of finding something new plummeted when we let the paleogelogists go. Then we let the seismics team go...that was literally flying blind.

So I'm trying to understand what you said as the discovery rate is not easily changed by oil economics because the exploration activity historically doesn't change much with falling prices?

jacob
Site Admin
Posts: 15907
Joined: Fri Jun 28, 2013 8:38 pm
Location: USA, Zone 5b, Koppen Dfa, Elev. 620ft, Walkscore 77
Contact:

Re: Too late to buy?

Post by jacob »

If prices are high, there is certainly an incentive to go out and drill new wildcat wells, ditto for any other resource. From a geological point of view, that is, estimating how much oil has been found though, any subsequent update in recoverable reserves is attributed to the original discovery of that field. It is not as if we can claim that "we found more oil" just because our initial estimates of how much oil was in that field we drilled 10 years ago were wrong. IOW, backdating allows for uncertainty in the estimates.

However, oil companies have traditionally claimed upward revisions as a new discovery. In the US, initial claims were given as P90 which made it very likely that subsequent upward revisions would appear giving the illusion of a continuous increase in reserves. Conversely, the Soviets reported P10 making it look as if they had a lot more reserves than they ultimately turned out to.

These preferences could perhaps not surprisingly be attributed to the various incentives in capitalism and communism respectively. Of course the unbiased approach would be to use P50.

Now, drilling wildcats is a process of diminishing marginal returns. Oil field size is somewhat distributed according to Zipf's law. Consequentially, large fields will most likely be found first (because they're easier to find). If you backdate, the consequence is that creaming curves will rise fast (early on, many years ago), and then taper off asymptotically as fewer and fewer new fields are found and as those fields additionally get smaller and smaller.

Today you can go out a drill like crazy ... but it's unlikely to lead to much new discovery because ... people have already gone drilling like crazy. Therefore the discovery rate doesn't matter much anymore (except of course to the economics of the individual company ... here a discovery of 50 million barrels matters because that's a lot of money. However on a global basis, it's less than a week's use)

https://www.google.com/search?q=creamin ... 37&bih=814

Looking at these curves, it's apparent that that the convergence is asymptotic to a constant ... which when you have new discoveries on the year-axis means that they're aren't finding much new oil at all.

Here the backdated discovery rate is directly (the time derivative of the creaming curve)
http://www.wolfatthedoor.org.uk/chartpa ... disav.html

Technologies like enhanced production techniques can raise the entire creaming curve, however. Production technology is rather mature, however, so what can be raised has already been raised. Again, this is a matter of convergence and adjusting the data.

In terms of prediction what happened was that initially, the world peak was predicted N years ahead. However, as technologies got better, N was pushed out. However, it wasn't pushed out as fast as time was passing. IOW, the time between the predicted peak year relative to the present year converged on zero. Technology was not enough.

Conversely, if prices are low, there's very little incentive to go out and find new fields because there would be less money to be made. This is why exploration activity is so sensitive to price.

TL;DR ---

I think perhaps the confusion is due to different ideas of what "discovery rate" means.

To geologists, discovery rate rate would be the rate of which new oil wells as discovered. This certainly varies by price as explores get fired or hired depending on the price. However, this discovery rate has very little influence on the shape of the creaming curve and it doesn't affect the total reserve number that it asymptotically converges to. It is that number which is important to predict the peak date.

To accountants and engineers, discovery rate would be the rate of which extractable oil in new AND EXISTING fields is revised upwards. They're not thinking in oil WELLS but only in producable oil.

Still TL;DR ---

Geologists are concerned with how quickly they are finding wallets lying around on the ground. If they find fewer and fewer wallets, they get concerned and can use stats to predict how many wallets can ultimately be found. On a finite planet, there's only a finite number of wallets.

Accountants and engineers are concerned with how much money they can dig out of the wallets that have already been found. However, just because they find that a wallet contains more money than they initially thought, it doesn't mean that they just discovered more wallets.

In fact, if you ask the geologists how many wallets have been found and how many can ultimately be found and you get a running update of P50 estimate from the engineers for each wallet, then you can calculate the total amount of money that you're likely to find, ever, without bias, because it's equally likely to be revised upwards and downwards. This total number is INDEPENDENT of the rate at which accountants and engineers "discover" money in the wallets.

NPV
Posts: 188
Joined: Fri Dec 06, 2013 9:41 am

Re: Too late to buy?

Post by NPV »

7Wannabe5 wrote:You can buy enough land on which to grow enough food to feed two people for $500 in and around Detroit. Come join me. It's really not that scary once you get used to it and it is more fun than living out in some dreary rural location.
How easy is it to select one which does not have a baggage of, e.g., utilities or tax debt? Also do you have to live there to own it?

7Wannabe5
Posts: 9369
Joined: Fri Oct 18, 2013 9:03 am

Re: Too late to buy?

Post by 7Wannabe5 »

I haven't heard of any problems with utilities and most abandoned property is purchased through tax auction. You don't have to live here to own property but you will get hit with stiff penalties and tickets and possibility of 90 days in jail for not engaging in regular maintenance such as snow shoveling and mowing. Property tax rates are also fairly high but just don't end up amounting to very much since SEV is so low (although likely higher than purchase price.) I expect my property tax bill this year on a quarter acre to be only about twice what I would otherwise have had to pay for annual rental on a very small community garden plot in an affluent suburb.

Post Reply