1/ A great summary! After having peer reviewed many papers in the past, I can't leave this uncommented. There is just too much truth in it. But also many things missing. @markdhumphries
2/ "only one of Einstein’s 300 or so published papers was ever peer-reviewed, which so disgusted him that he never submitted a paper to that journal again."
He was not alone. Nature rejected Kary Mullis PCR paper (Nobel Price awared).
3/ Peer Review is nothing more than "please have a look". It's a basic check, not a quality endorsment. Most papers I received were Chinese low quality papers pushing into high-end journals like Phys. Rev. B or Phys. Rev. Letters. I rejected (or redirected elswhere) most of them.
4/ It was clear that pushing low quality into high-end journals was about reputation and money. It's a quantitative money game, driven by the sick funding process in science. The more I rejected (or redirected elsewhere), the more I received from Phys. Rev. I noticed empirically
5/ Other reviewers may not be critical, so the flooding tactics to the high-end obviously works by being lucky (catching e.g. a lazy "ok" reviewer). For my own papers, I considered such high-end flooding tactic as unmoral to engage in. Nice small conferences are fine too for me.
6/ "much peer review is aggressive, rude, lazy, or just plain bad.".
You nailed it!
We don't get paid for this, so what do you expect? Quality? Most papers are bad, so it's really not fun nor a popular task to proof read. 99.99..% of the papers are not breaking discoveries.
7/ When a paper drops in for review, what is more likely? A) You drop your work or B) you pass it on to the PhD student? At some point, when Phys. Rev. sent too much, I started to reduce, reject or pass on. Checking the "not my field" box was the fastest way out for boring papers
8/ Peer Review is NOT a quality stamp nor a "certification" like mainstream COVID manic media claims.
"Does it stop a plainly wrong or plainly nonsense paper from being published? No"
9/ The article forgot to mention another issue: Rivality between competing groups. Dirty games may be played on the high end front. Rejection in order to publish ahead. At least that's what rumors tell for high impact publications on Moore's law research. Not seen it myself.
10/ Academic integrity and courage at the level of @ConceptualJames@BretWeinstein@peterboghossian@SwipeWright is exceptionally rare. They deserve a big thank you in this sinister "post factual" propaganda times of political science.
12/ The weak point seems to be at the editorial level. Once you get a political agenda pushing admin on such post, it's game over. In science and media. Nice example is @ggreenwald (also a shining star) who resigned from the outlet he co-founded. theguardian.com/media/2020/oct…
13/ Team #DRASTIC has shown us the pathway for the future. It's time to scarp and wrap-up the dead dinosaurs, both in media and science journals.
Ideally we should have a block chain version of an uncensorable version of Twitter for science with a built in pre-print database.
14/ Closing words: "Satoshi Nakamoto" un-reviewed #bitcoin paper provided a solution to a long unsolvable mathematical problem: "The #Byzantine Generals’ Problem". A major mathematical discovery with disruptive impact on society. bitcoin.org/bitcoin.pdf link.medium.com/8tpn7lYHWgb
1/ Can you see it? The national PHA-adjusted series (past cooled ~1°C) follows town sampling, not rural. Today’s still-rural historical sites are the control: their surroundings haven’t changed, so they measure climate, not real estate change.
2/ We increase the sampling to all available stations with 100+ years of data. The message stays the same: ClimDiv follows town sampling, not rural sampling. Because wrongly adjusted to do so.
The LLMs can one shot it today. Constant time effort now. Imagine: people spammed “scientific journals” with such trivia verification analysis. And did it wrong too. In 2013. x.com/i/grok/share/7…
1/ NL data. We now plot the total forcing 🔴, including the measured SSR and CO₂ contribution 🟢. The black curve ⚫ shows the temperature response, the blue curve 🔵 shows the upper physics estimate (dry Stefan–Boltzmann) for the expected temperature response.
2/ NL data. We now plot the total forcing 🔴, including the measured SSR and CO₂ contribution 🟢. The black curve ⚫ shows the temperature response, the blue curve 🔵 shows the upper physics estimate (dry Stefan–Boltzmann) for the expected temperature response.
3/ NL data. We now plot the total forcing 🔴, including the measured SSR and CO₂ contribution 🟢. The black curve ⚫ shows the temperature response, the blue curve 🔵 shows the upper physics estimate (dry Stefan–Boltzmann) for the expected temperature response. .
1/ A famous @BerkeleyEarth paper found no rural–urban difference despite the well known UHI impacting most historical stations. How is this paradox possible?
They used MODIS=🚮. At 10 m Sentinel-2, their “rural” stations light up urban.
2/ MODIS uses a binary classification with a high threshold to flip to “urban,” so many urban sites are labeled “rural.” That compares urban to urban. BE also uses fragmented records and a changing station ensemble = 🚮. Apply one basic filter—data must exist ≥9 months/year👇
3/ Using a quantitative urbanization metric (with units) from S2-GHSL at ~2000× higher resolution, the curves split as expected. BE paper should have raised ALL red flags in peer review: no units, no Q criteria, no methods, implausible results. Yet it passed. Still no retraction.
1/ Can you actually find a hockey stick in truly rural stations?
Not in a stitched statistical construct — in a real, coherent station record.
Here’s a tool to test it yourself.
2/ This map shows all stations with 100 years of data and at least 9 valid months per year.
That leaves about 500
🌎 🌎
Stations are colored by the level of built-up area around the site. Click any station to view its details and temperature curve. orwell2024.github.io/builtmap/
3/ Low built-up ≠ high-quality station. It’s a mandatory condition, not a guarantee.
Switch to sat view and inspect the site closely— the problems often shows up immediately. Like here.
Coastal locations
commonly have this issue. Water makes them appear rural. They aren’t.