Continuing the last article’s theme of ‘too good to be true,’ we now turn our attention to LK-99.
LK-99 is a compound that researchers in South Korea claimed could serve as a room-temperature ambient-pressure superconductor.
Superconductors are materials that, when cooled to a “critical temperature,” conduct electricity without energy loss. The most common use of superconductors today is in MRI machines, but future applications include quantum computing, digital circuits, and power transmission. Unfortunately, superconductors need to be kept below the “critical temperature,” where the current warmest of these is -209.5°F. One experiment has managed room-temperature superconducting, but with a catch: the researchers increased pressure to over 2.6 million times the atmospheric pressure at sea level on Earth.
The cost for cooling or pressurization makes most superconductivity cost prohibitive. A room-temperature ambient-pressure superconductor would be a game changer. And along came LK-99 to both great fanfare and skepticism.
“Excitement rises over possibly revolutionary ‘miracle material’ – but there is still no good reason to believe it exists”
The Independent
LK-99 Foomscrolling
The hype and enthusiasm around LK-99 are frankly more interesting than the technology (even if it’s real.) The term “foomscrolling” was coined to describe the public’s insatiable appetite on social media for stories about new groundbreaking technology like LK-99 or ChatGPT.
This rise and fall of enthusiasm, or “foomscrolling” as it’s been dubbed, is not unique to LK-99. It often occurs when emerging technologies can be visually demonstrated and have the potential to profoundly impact humanity.
Fagen Wasanni Technologies
And more than the technology itself, the levitating rock seems to have captured the public’s imagination. The original, questionable experiment even spawned a series of fake online videos.
Unfortunately, and unsurprisingly, the LK-99 hype was short-lived. Immediately there were red flags, like one of the authors citing “many defects” or the lack of a heat anomaly, and very quickly the original findings failed to replicate.
LK-99 Probabilities & Prices
Below is a summary of available platforms and their forecasts on the chances of the LK-99 paper being replicated or confirmed as real (as of August 14at 9:00 p.m. ET):
| Platform | Description | “Yes” Criteria | Close Date | Peak Yes | Current Yes | Volume |
| Polymarket | Real money prediction market (prohibited in US) | Paper findings are “replicated at least two times by independent researchers unaffiliated with the authors of this paper” | 12/31/2023 | $0.56 | $0.10 | $2,595,899 bet among 894 traders |
| Kalshi | Federally regulated exchange | “If an eligible peer-reviewed journal publishes research” confirming superconductivity | 12/31/2024 | $0.43 | $0.11 | 160,626 |
| Insight Prediction | Real money prediction market (prohibited in US) | “replicated by at least two other independent research groups that share no affiliation with the authors of this paper” or will resolve NO before expiration if “the paper(s) are retracted from Arxiv…and a consensus of credible reporting determines that the results were fraudulent or otherwise fabricated” | 12/31/2023 | $0.432 | $0.08 | $19,441 |
| Good Judgement Open* | Crowd forecasting site | “When will a major scientific journal publish a peer-reviewed article confirming that the material LK-99 is a room-temperature superconductor working at ambient pressure?” | 12/31/2024 | 37% | 12% | 81 forecasts by 47 forecasters |
| Manifold** | Play money prediction market | “Will the LK-99 room temp, ambient pressure superconductivity pre-print replicate before 2025?” [Peer review not necessary] | 12/31/2024 | 66% | 12% | 5,483 traders |
| Metaculus** | Forecasting platform | “Will any peer-reviewed replication attempt before 2025 confirm the discovery of room-temperature and ambient-pressure superconductivity in LK-99?” | 12/31/2024 | 37% | 1% | 1,266 predictions by 414 predictors |
| Futuur (real money) | Real money prediction market (prohibited in US) | “successful replication is achieved by the end of 2024” with additional criteria regarding temperature, pressure, and Meissner effect. Compound must be correctly synthesized. Study must be independent of the original. Peer review not necessary and “pre-prints can be taken into account.” | 12/31/2024 | 68% | 22% | $320.32 among 19 traders |
| Futuur (play money) | Play money prediction market arm of the real money platform (play money is open to US) | “successful replication is achieved by the end of 2024” with additional criteria regarding temperature, pressure, and Meissner effect. Compound must be correctly synthesized. Study must be independent of the original. Peer review not necessary and “pre-prints can be taken into account.” | 12/31/2024 | 91% | 9% | 15 traders |
| Zeitgeist | Real money prediction market (prohibited in US) | “if the material is generally agreed to be a superconductor at temperatures at or above 15 degrees Celsius (~ 59 F) and at or under 1 atmosphere of pressure (1’013 millibars). It must be replicated by at least one independent researcher unaffiliated with the authors of the paper” | 12/30/2023 | 43% | 28% | 6 traders |
*Good Judge Open’s question uses four banded date ranges. We’re treating the last range, “Not before 1 January 2025” as “No” and the other three answers as “Yes.”
**The table above only includes the peak Yes price after the first 24 hours. Manifold (90%) and Metaculus (89%) Yes forecasts both spiked within the first day of the market/question creation, but are not used in the table above.
Two observations about the “peak” probabilities and prices of the various LK-99 markets:
1. In looking at when each market’s (or forecasting platform) Yes price/probability peaked, it worked out the opposite of what we expected. The platforms with real money and high volume tended to peak after the platforms with play money, low volume, or relatively new to the space. Here are the dates of their respective peak Yes prices:
- Metaculus on August 6
- Polymarket, Kalshi, and Good Judgement on August 5.
- Zeitgeist initially started at 50-50 and peaked at 43% for the first five days from August 2-6.
- Futuur (real money) at 68% on both August 1 and 3, while Futuur (fake money) on August 1.
- Insight on August 2
- Manifold the night of July 31-August 1
2. The platforms with the highest peak Yes prices were either real money with extremely low volume (19 traders) or fake money:
- Futuur (play money): 91%
- Futuur (real money): 68%
- Manifold (play money): 66%
Polymarket was the only other platform to crack 50% ($0.56/share) despite being real money and high volume.
Among the remaining sites, Yes prices on Kalshi, Good Judgement, Insight, Metaculus, and Zeitgeist all peaked between roughly 37-43% (if we treat prices as the implied probabilities). They are all either real money markets or forecasting platforms.
Why bet Yes?
Given all the negative information that has been coming out about the researchers and original papers, the probabilities and share prices for Yes still seem too optimistic. It’s not clear how much the remaining Yes shares are holding out due to truly believing in LK-99 chances at superconductivity or are merely holding out hope for scientific progress. This could be the classic case where people are betting on what they want to happen rather than what they think will happen:
Congrats to all of the bettors in the superconductor @Polymarket who rooted against progress. Enjoy your money, you cynical sons of bitches.
— Gaeten Dugas (@GaetenD) August 9, 2023
For more information about forecasts and prediction markets related to LK-99, check out:
- Base Rate Times is currently under maintenance, but their X account is posting updates on the replication and retraction of the LK-99 initial paper.
- Metaculus has a page devoted to all its superconductor questions.
- Manifold has a page with related markets.