Examining flawed approaches in bitcoin energy research
Source: bitcoinmagazine.com
Accurate data forms the foundation of any credible analysis, especially regarding something as intricate and debatable as bitcoin mining. Many of these studies, such as the one from Chamanara et al., often depend on outdated or incomplete information, leading to conclusions that are not merely misleading but potentially detrimental when utilized to shape policy or public perception. In the realm of bitcoin mining, where the landscape continuously evolves due to technological innovations, regulatory shifts, and market changes, relying on stale data equates to trying to navigate with a broken compass.
Consider the reliance on the Cambridge Bitcoin Energy Consumption Index (CBECI) data in the Chamanara et al. study. While CBECI is one of the more frequently referenced sources for bitcoin energy consumption, it is not without its shortcomings. The index itself relies on a model that estimates energy use by combining hashrate data and assumptions about mining hardware efficiency. However, as pointed out by Lei et al. (2021) and Sai and Vranken (2023), the model has considerable limitations, especially regarding the rapid advancements in mining hardware efficiency and the geographical movement of mining operations.
My sacred altar, which I maintain in my closet, is a meticulously carved, sophisticated yet austere dedication to Koomey, Masanet, and Shehabi for their decades of effort to enhance data center energy modeling. These masters of computation have made it abundantly clear to me: if you lack bottom-up data and depend on past trends while disregarding IT device energy efficiency trends and what drives demand, then your research is nonsense. Consequently, with one sweeping yet very precise action, I dismiss Mora et al. (2018), deVries (2018, 2019, 2020, 2021, 2022, and 2023), Stoll et al. (2019), Gallersdorfer et al. (2020), Chamanara et al. (2023), and all the others cited in Sai and Vranken’s thorough literature review. May these perish in one explosive yet metaphorically grand mega-fire somewhere off the shores of the Pacific Northwest. Journalists and decision-makers, I beseech you to cease consulting Earthjustice, Sierra Club, and Greenpeace, for they are unaware of their misdeeds. Forgive them their transgressions, for they are merely followers. Amen.
For far too long, we have been subjected to the repercussions of deficient academic studies on the energy consumption and environmental repercussions of bitcoin mining. The results of this inadequate research have led to alarming news headlines, transforming well-intentioned individuals into furious lawmakers and irrational advocates. To spare you the agony of encountering one of these careless papers, I have devoted myself to the bitcoin mining deities and conducted a comprehensive examination of a publication from the United Nations University, recently released in the American Geophysical Union’s Earth’s Future. Only the most fearless and dedicated bitcoin aficionados may continue reading; the rest of you may return to observing the price fluctuations.
As I began to jot down notes on the document, it became clear that Chamanara et al.’s work was quite perplexing. The study puzzled me because it’s a poorly constructed piece that hinges its existence solely on de Vries and Mora et al. It employs data from the Cambridge Center for Alternative Finance (CCAF) Cambridge Bitcoin Energy Consumption Index (CBECI) without acknowledging the model’s shortcomings (refer to Lei et al. 2021 and Sai and Vranken 2023 for a thorough critique of the issues surrounding CBECI’s modeling). The authors blend their results from the 2020-2021 timeframe with the reality of bitcoin mining in 2022 and 2023. They also leaned on some environmental footprint methodology that might lead one to believe it was feasible to diminish or expand a reservoir based on how much you engage in Netflix and chill. In fact, this is what Obringer et al. (2020) inferentially suggest is feasible, and the UN study cites Obringer as one of its methodological pillars. For the record, Koomey and Masanet were not fans of Obringer et al.’s methodology either. I’ll light another soy-based candle at the altar in their honor.
Another major concern is the mixing of data across several years. By aggregating energy consumption figures from 2020 and 2021 into a single statistic, the authors obscure essential changes that transpired in the bitcoin mining sector during that timeframe. For instance, the Chinese government’s crackdown on bitcoin mining in 2021 led to a significant departure of miners from the nation, resulting in a marked redistribution of hashrate to other areas, including the United States and Kazakhstan. This transition had a profound effect on the energy composition and efficiency of the global bitcoin mining network, yet the Chamanara et al. study neglects these dynamics, instead portraying a static and outdated picture of the industry.
- The authors combined electricity usage over several years, overextending the potential revelations of their findings based on their methods.
- The authors depended on historical trends to formulate current and future suggestions despite abundant peer-reviewed literature clearly illustrating that this can lead to overestimations and exaggerated assertions.
- The study promises an energy computation that will clarify bitcoin’s actual energy consumption and environmental impact. They utilize two datasets from CBECI: i) total monthly energy consumption and ii) average hashrate share for the top ten nations where bitcoin mining is conducted. Remember, CBECI is based on IP addresses tracked at various mining pools. CBECI-associated mining pools constitute an average of 34.8% of the total network hashrate. Thus, the data employed likely comes with considerable uncertainty.
The significance of precise data in bitcoin mining assessment
Ultimately, the importance of accurate data in bitcoin mining analysis cannot be emphasized enough. Without it, we risk making decisions based on flawed assumptions and obsolete information, with potentially severe implications for both the industry and the wider public. As the saying goes, “garbage in, garbage out.” If we hope to engage in a meaningful dialogue regarding bitcoin’s energy consumption and environmental effects, we must begin with trustworthy, up-to-date data and a rigorous, transparent approach. Anything less is mere noise.
Having set the stage for you, my devout reader, I will now share a tale about a recent bitcoin energy research effort. I pray to the bitcoin deities that this will be the last one I pen and the final one you’ll ever need to peruse, but I sense that the gods are punitive and shall spare no mercy on my spirit—even during a bull market. One deep breath (cue Heath Ledger’s Joker) and here… we… go.
Additionally, the CBECI data referenced in the Chamanara et al. study originated from an earlier version of the model, which has since been revised to better reflect accurate machine efficiency metrics. This suggests that the energy consumption figures presented in the research are likely inflated, distorting the findings. The fact that the authors did not revise their analysis to incorporate the latest data is a significant oversight that undermines the validity of their conclusions.
Here’s a more explicitly articulated summary of the core issues with Chamanara et al. (and by the way, their lead author never replied to my email seeking their data so I could, you know, verify, not trust. 🥴):
Your sensitive ears might have recoiled with disbelief at my bold assertion in the introductory statement that the most prominent research on bitcoin mining is nonsense. If you’ve ever perused Jonathan Koomey’s 2018 blog post on the Digiconomist—also recognized as Alex deVries—or his 2019 Coincenter report, or Lei et al. 2021, Sai and Vranken 2023, Masanet et al. 2021, or… Well, the gist is that there are thousands of words already articulated demonstrating that bitcoin mining energy modeling is in disarray and that this is not unique to bitcoin! It’s a challenge that data center energy studies have wrestled with for many years. Experts like Jonathan Koomey, Eric Masanet, Arman Shehabi, and the lovely Sai and Vranken (sorry, we’re not yet on a first-name basis) have penned enough material to probably cover the walls of at least one men’s restroom at every bitcoin conference held last year, confirming this reality.
On a slightly grim October afternoon, I was tagged on Twitter/X regarding a new study on bitcoin energy consumption from authors linked to the United Nations University (Chamanara et al., 2023). Little did I anticipate that this study would provoke my intense focus to such an extent that I would slip into my own unique drug-induced-gonzo-fear-and-loathing-in-Vegas state, becoming hyper-fixated on this research for the subsequent four weeks. While I may be exaggerating about the intense substance use, my memories from this period feel like a technicolor, fevered dream akin to a toxic relationship. Remember Frank from the critically acclaimed 2001 film, Donnie Darko? Yep, he made an appearance too.
Conversely, a more accurate evaluation would entail a year-by-year analysis of energy use, considering the changing efficiency of mining hardware, the geographical redistribution of hashrate, and the implications of regulatory changes. This method would offer a more nuanced and precise comprehension of bitcoin’s energy consumption, facilitating more informed discussions about its environmental repercussions and the potential for future enhancements.
