Never noticed News.YC's CO2Stats before. How do they compute this stuff anyway? | Hacker News Reader

Never noticed News.YC's CO2Stats before. How do they compute this stuff anyway?

co2stats.com

27 pointspius60 comments

Smoke, mirrors, reading of dead chicken entrails...

You know, the usual PT Barnum stuff to impress the rubes.

Interesting paragraph from their site: "CO2Stats reserves the right to cap REC purchases in case of excessive use, as its own discretion."

That is, you could pay them $100 a month and they may or may not purchase offsets to cover your 5 million page views.

Thanks for your comment! In fact, every time our customers' footprints have gone over their monthly cost, we've paid for the full corresponding number of RECs. That clause is there to protect against gross abuse. For example, if a site consistently receives 20M hits per month (a custom tier) and signs up for a 5M hit per month package, we reserve the right to manage only the first 5M.

Ah, so the clause is specifically meant to exclude the case that a website buys RECs for less than they actually use? That's perfectly reasonable, but I have to admit when I read the bit above, I didn't interpret it that way. You might want to think about tweaking the language to make it harder for skeptics to dismiss you.

No measurement of this type is ever going to be perfect. Do you think, because of that, that it's not even worth trying?

I can see why CO2stats may be successful, phrases like "purchases renewable energy to neutralize it, so your visitors feel good about staying longer." are quite compelling for some people...

Ah this website uses a ton of power. CO2 is evil. CO2Stats "neutralizes" it so that it's all fine.

Wouldn't it be a better idea to help companies reduce their power usage, rather than "buying renewable energy" to offset it? Or encouraging them to install their own renewable energy sources.

If I murder 2 people, and offset it by having 2 children, is that fine? :/ (As long as I display a button on my T-Shirt saying I'm "Person neutral")?.

Given the wild inaccuracy of geo-ip lookups and the million other factors at play, I don't see any value in guessing at numbers here...

There's value in public steps taken to be green (engaging, marketing) and there's value in saving energy. There are externalities which can't be accounted for by the process of simply reducing consumption, and offsetting, by RECs or otherwise, should be considered something to do after conservation and reduction, but a valid last step.

If anyone should try it's this guy - there are so many cowboys running amok in the space that someone with this sort of background is, for better or worse, going to pass the green "sniff test" and be above suspicion. At least to the extent that FUD allows.

yes

Down-modded again. hmmmmm

Too terse? Or just a way of anonymously trashing a view you don't espouse?

I know people get tired of comments about moderation. I get tired of the stupid moderation system. Maybe at some point the frustration on both sides will be addressed.

There's no quicker way to get down-modded than to talk about getting down-modded.

I can speculate that the reason you were down-modded was that you didn't explain why it's not worth trying?

Yes. I realize I stepped in it.

I hope you are right. I was getting ready to make a long argument against attempting to measure things with so many unknowns, but then folks tell me I'm long-winded.

So I figured I'd just give a simple answer and if anybody had a question I would be glad to explain myself. I was assuming that a reader would not use the down arrow as a way of disagreeing and would instead simply ask "Why?"

Didn't work out that way, though.

Obviously I can't speak for others, but I'd prefer the long winded response over the too short response. Or maybe try and find the happy medium.

The reality is, that all proposed "climate change" models are mathematically suspect and thus there is no actual evidence that reducing CO2 from man-made sources would in fact modify global temperature.

For instance, sunspot activity or the lack of it, has a far greater correlation to short term warming or cooling.

For longer term warming or cooling, any true scientist should readily admit that the tools currently available have very limited utility, since the mathematical models cannot possibly take into account enough of the real behavior.

For example, the Gulf Stream, which we know has a great deal of influence on East Coast weather, is one of the most studied, yet still is not understood very well; thus no mathematical model of its behavior can be said to be accurate.

CO2stats is more about selling "liberal guilt credits" than anything else.

Throwing rocks at a hornet's nest ;)

> there is no actual evidence that reducing CO2 from man-made sources would in fact modify global temperature.

There is no evidence that such a modification in temperature would be a bad thing, either.

Basically, it's just a stack of assumptions. I'm not saying it's necessarily false or anything, but I find it strange to see supposedly intelligent, rational-minded, etc. people (= "hackers") step into this one as easily. Let's question all those assumptions first, would you?

pg, let's ask a different question. Measurements aside, is what they are offering really reducing the carbon footprint of a website?

I would argue no, it does not. And here's why: They are buying credits which make clean energy cheaper. The trickle down effect (not a reference to Reagan) means that others will be buying cheap energy, which is good, but it does not mean you will be buying cheap energy. So I don't see this as directly translating into reducing your carbon footprint. Moreover, who actually gets the reduction in their personal carbon footprint here: the person actually using the green energy or you, who helped pay for this energy by proxy. I would argue that it is to a much greater degree the person who actually bought this energy who gets most of the benefit. You would be much better off using your time to plant trees in proportion to the amount your site uses or finding a green hosting service than use CO2Stats.

Even if CO2Stats is measuring this even remotely accurately (which they couldn't possibly be), their model of buying these renewable energy credits still translates into a very low reduction in your actual carbon footprint.

Having said that... is their service worth whatever it costs? Most likely no, since for it to work you would have to pay a disproportionate amount to get an amount that would indeed offset your footprint. Is CO2Stats capitalizing off people's supererogatory desire to do good by nature, regardless of whether there are no benefits? You bet.

You can't argue that they aren't doing good for the environment. At least they're making a small impact, even if it's not a total "offset" of the internet experience. In fact, they don't even claim to offset it; they claim to "neutralize" it with clean energy in proportionate amounts.

Actually increasing the supply of alternative energy could easily increase the demand for energy as energy prices decrease. There could be net greater CO2 output.

CO2 offsets are bullshit. A more reasonable solution is to fund research into alternatives that are economically viable without subsidies.

Tracking energy used, as CO2Stats does, is still an important part of the solution. But buying offsets doesn't necessarily mean you're doing good.

On multiple levels, they are like buying an indulgence.

There will always be some amount of demand for energy. If you pump a little bit more wind powered energy into the grid that reduces the amount of demand on coal, then you're doing at least a little bit of good.

That's not the way supply and demand work.

Adding wind increases the supply and lowers the cost of energy, which increases the demand. It could easily cause a net increase in energy used.

The fundamental problem is that you don't offset anything. Offsetting would be replacing conventional energy generation with alternative, and paying for the difference in cost, not just adding more energy.

Your local energy provider probably has a program to do this. Call them up, and ask to spend more on energy. As they expand their generation capabilities, they do more alternatives according to your direction. That's a "real" offset.

I think the certificates that CO2Stats buys actually do what you described as a "real" offset. It entices traditional energy suppliers to invest in clean energy.

I'll look into it. If so, awesome!

4. How do Renewable Energy Certificates help the environment? Renewable Energy Certificates guarantee that a certain amount of clean wind and solar electricity will be generated onto the power grid, displacing fossil fuel-based electricity generation. Since we measure the location and fuel makeup of your users and servers, we are able to purchase an equal amount of renewable energy to offset the non-renewable component of your energy footprint. The more renewable energy we support, the less non-renewable energy will be required in the future.

http://www.co2stats.com/faq.php

I'm still a bit unclear on whose behalf they are buying energy.

It is worth trying if you're trying your smartest.

If you're simply abdicating the responsibility of doing the hard thinking to someone else, then you're not trying, just flaunting a green badge.

This type of measurement ought to be perfect.

1. Computers can report how much power they are using. Laptops directly, for servers you might have to do something fancy like keep track of uptime and check your power supply's wattage.

I think the latest Laptop CPUs can reduce power consumption when not fully loaded, not sure. But I don't think server chips are that complicated... yet.

2. Call your power company and ask them how much of the power they provide is hydro, coal, wind, solar, nuclear, etc.

3. Do some math.

That should give you exactly how much CO2 you're using.

The next question is, does funding green energy sources make sense?

It might not. It might be worse then doing nothing. Here's how:

High tech has a much lower CO2/output ratio then almost any other industry. So it would be much easier for high tech to pay a green penalty, especially if it comes with bragging rights.

However, the other 90% of CO2 producers, will only switch if it is profitable.

And here's the crux, by funding unprofitable green sources, you could be hurting the development of ones which are truly economical.

OPEC is cutting production with oil over $100 a barrel, for sweet zombie jesus' sake! Alternative energy ought to be competitive without help.

If you're still worried about your CO2 emissions, pay someone to plant a tree, that won't compete with green tech. Or buy more stuff from an organic farmer, because organically worked soil is a huge CO2 sponge. Or spend some money on lobbying politicians.

Or at least don't give money to someone who: " reserves the right to cap REC purchases in case of excessive use, as its own discretion."

I guess mirroring the OP, I like this concept behind C02Stats, but I am very curious as to how this is precisely calculated (or even roughly estimated.) Do CO2Stats know the power grid the servers are on? How do they account for the disparity in power usage behind different systems (a mainframe != laptop)? Without having detailed knowledge of a building, how could they possibly know about air conditioning costs or other secondary electricity use (lights, etc)?

Averages?

Great question! We measure quite a variety of signals, including but not limited to client locations (e.g., is traffic coming from coal-burning areas?), visit lengths, server locations, transfer sizes, transfer times, and window sizes. These signals are fed into a "hand-built" database that allows us to compute aggregate carbon footprints.

Very interesting . . . best of luck with the venture!

Thanks!

I'm curious: Why do you keep sending posts to your server every few minutes?

The CO2Stats calculation includes the contribution of client electricity use. In order to more accurately measure the time clients spend on a given CO2Stats-powered page, it's necessary to have the client infrequently check in.

Do you include the CO2 stats overhead (the pinging, etc) in your calculations for neutralizing?

Absolutely!

I thought I heard the idea for CO2stats somewhere before...

http://en.wikipedia.org/wiki/Indulgence

Oh yeah, that's it.

It's not quite as bad as that. It's better to burn coal and buy a few trees than it is to just burn the coal, right?

While I don't think that carbon offsetting is going to solve any long-term problems, it is nice to plant some trees and have a bit more money poured into renewable energy research.

We're not going to see real change until we are out of coal and oil. Hopefully that will happen soon. (Yes, it will be painful. But the sooner the pain starts, the sooner it will be over with.)

You should cut your both arms then, so you won't have to suffer losing them should you become involved in an accident.

That is a different situation. It's not likely that I will ever lose my limbs. It is likely that we will run out of oil and coal some day, since we use a lot and there isn't a lot left.

I think you may be mistaken on ther enot being a lot of coal left:

"According to the widely accepted view, at current production levels proven coal reserves will last 155 years (this according to the World Coal Institute). The US Department of Energy (USDoE) projects annual global coal consumption to grow 2.5 per cent a year through 2030, by which time world consumption will be nearly double that of today."

From: http://www.energybulletin.net/node/29919

The article goes on to say that we probably don't have 155 years of coal reserves, but it should still last for some time now, unless the rate of usage sky rockets.

So your argument is that the human race will be wiped out in 155 years, so we don't need to worry?

No, my argument would be that in 155 years we'll have better alternative sources of energy. Wind and Solar are already viable alternatives, with time their cost will come down and their efficiency will go up potentially eliminating the need for fossil fuels such as coal.

Oil will get scarce, soon. http://en.wikipedia.org/wiki/Coal#World_coal_reserves says:

"At the current production rate, [coal reserves] would last 164 years."

Thanks. One more regarding certificates: My guess is that a certificate subsidizes the cost of more expensive clean energy so that it equals the cost of non-renewable energy. Is this right?

Either way, very interesting business model. Best of luck!

Thanks, Mwerty! That's right, RECs function subsidize the difference between renewable/non-renewable sources.

So if a geo-ip lookup on my location (Which is extremely inaccurate) says I live near a coal station, you assume that I use that power and not power from some other source?

I guess they are just taking the average power mix of your country/state.

Hey CO2Stats - A couple questions for you:

How do you account for different monitor sizes, types and brands and poor wasteful electricity in older buildings?

Thanks!

There is no way they could account for the types and brands of the monitor, the efficiency of the individual computers, etc.

They also cannot make an accurate assumption about where the power is coming from because there is such a thing as an electric grid, which draws power from a variety of sources.

Furthermore, the calculations to determine what goes in to processing a page have to be almost comical. For example, if it was Google, one search request can take thousands of machines, not to mention all the crawling required to produce those pages. Good luck getting that information.

There are so many holes that need to be filled here that their calculations cannot even be remotely accurate.

CO2Stats is bogus. I hope they don't charge anybody, because that would be fraud.

Perhaps they should publish their algorithm, to provide a little transparency and openness.

You ever notice that whenever you walk into a 7/11 at 2am there's always at least one carton of orange juice there waiting for you, even if you've never been there before? I think you're missing the forest for the trees, so to speak.

I understand what they are trying to do. But the sheer inaccuracy of what they provide clearly negates the supposed benefits.

Would you donate to a charity that mismanaged or wasted most of your money, if only just because donating to charities is good? Or wouldn't that mismanagement negate most of what is good about donating to that particular charity.

I'm sure the people behind CO2Stats are not stupid. So they must be aware of the shortcomings and limitations of their service. Yet they continue to run the service anyways, calling it accurate. That is dishonest.

(and no, questioning CO2Stats does not make me a conspiracy theorist)

"But the sheer inaccuracy of what they provide clearly negates the supposed benefits."

How? Do you understand what it is that they do?

People pay them to offset the carbon footprint of their site. Except if they are paying them an amount based on an inaccurately calculated footprint, that's a problem.

If you wish to continue drinking the Kool-Aid, go ahead! It sure is tasty! But don't blame me when your teeth fall out.

Problems like this need to be solved. That is, the real world demands that we estimate things that cannot be estimated reliably. To take a less controversial example than global warming, consider the risk of avian flu. Clearly, we cannot get a reasonable quantitative estimate of the chance of a flu epidemic, or how many people it would kill, etc. Nevertheless, we are forced to decide what preparations (if any) to make anyway.

Just because the problem is hard, doesn't mean it isn't worth attempting. I see no problem with what they are doing, as long as they are honest about the uncertainty in the estimates they hand to you.

Putting the global warming debate aside, I totally agree with you.

However, I don't see them being transparent. Furthermore, I think given the complexity of determining what they claim, the simplicity of their model is almost certainly not adequate.

I like your Avian Flu example. I think in that case, we would both agree that there is a point at which preparations for such an event would be almost useless when there is a certain lack of information, when what we do know and what we don't know is so muddled that we are not even sure what direction to take. Obviously doctors and scientists are not blind here and know generally what direction to take in their preparations, just not to what scale.

What I am arguing in terms of CO2Stats here is that the likely inaccuracy of their calculations just about voids what they are offering because you don't know that what you are getting is in the proper proportion or even that it actually helps to the degree that they say it does.

There's a difference though between taking a best guess because we need to know, and selling a service based on your best guess without putting that methodology up for scrutiny.

Without seeing how they calculate your imprint, they're basically working on the same "trust us" principle as the guys that sell that Q-ray bracelet.

Do you?

I think that's his point.

Inspect the data that their "complex heuristics"-based widget sends back about the client. I'm positive that you'll be pleasantly surprised at the breadth of information.... or, it sends nearly nothing. And nothing of relevance to actual consumption. But c'mon, CO2Stats is totally legit, guys... really...

Interesting concept. What I don't understand is why as someone running a website I should offset the CO2 from my data center, my client's homes and the networks between them.

For the same reason that Amazon has free shipping?

Math.random()

I don't think the numbers matter. It's just a badge to say "We're green" isn't it? Seems like it's leveraging off the current "CO2 is evil" fad.

Instead of a seal that nobody cares for (hack-proof, privacy seals are OK, but green-certified seal?), why don't they enter the carbon credits market?

This seems like a misuse of technology. Use this technology to help companies trade carbon credits. A few bulge bracket banks are already in this game. They buy carbon credits from African cooperatives cheaply and sell them at a healthy profit to factories who need these credits.

And they are making millions from it. What are you waiting for guys?

Reply on news.ycombinator.com