If MakerDAO had the insurance fund mechanism covered in an earlier article at the time, could this incident's loss have been completely avoided?
Not completely avoided, but substantially mitigated. What an insurance fund can do is 'cover the bad debt gap' — that is, the outcome caused by this incident's third domino — but an insurance fund can't prevent the first two dominoes (macro panic causing price to crash, network congestion causing the liquidation auction to lose competition) from occurring — the structural causes triggering the cascade aren't a problem any single protocol's internal capital reserve can unilaterally solve.
An insurance fund's actual role is: if MakerDAO had already had a sufficiently sized insurance fund at the time, this $4M to $5.67M gap might not have needed to be covered through emergency additional issuance of governance tokens (an approach that dilutes all existing token holders' stake), instead resolved directly by deploying the insurance fund — a smoother way to handle it for both the protocol's overall long-term confidence and token holders' stake. This is also why the insurance fund mechanism gradually became a standard configuration in most lending protocols after the Black Thursday incident — it doesn't solve 'whether a cascade will happen,' but the question of 'who bears the gap once a cascade genuinely happens.'
Did the liquidator who 'bought $8 million worth of ether for $1' in this incident bear any responsibility or face any consequence afterward?
No, and this outcome itself is an important piece of understanding this incident. This liquidator's operation fully complied with the liquidation mechanism's rules at the time — they genuinely paid the bid amount the auction required (1 DAI), genuinely obtained the corresponding collateral through the protocol's designed process, with no exploitation of any code vulnerability or violation of any explicit rule throughout — purely because no other liquidator was competing against them at that moment, they won the bid at such a low price. From a technical and rules standpoint, this was a loophole-like outcome the mechanism's design itself produced under extreme conditions, not any individual's violation, and consequently no applicable accountability mechanism existed.
This is exactly one of this incident's most valuable lessons for the entire DeFi industry: when a system's correct operation depends on the assumption that 'there will always be enough participants competing against each other,' once this assumption fails under extreme conditions, the system itself produces an outcome that looks unfair yet is entirely compliant. This is also why most protocols afterward chose to improve the auction mechanism's technical design (such as a Dutch auction, batch auction, or introducing additional price protection mechanisms), rather than trying to hold individual participants accountable — because the problem's root lies in the mechanism design, not any individual's behavior.
What concrete changes have current lending protocols made to avoid a similar Black Thursday scenario recurring?
A few common concrete improvement directions: redesigning the auction mechanism, with most protocols introducing a more refined auction process, such as setting a minimum bidder count threshold, or switching to an auction model that runs over a period, giving more liquidators a chance to participate in bidding, rather than the original design potentially getting won instantly by a single liquidator at an extreme low price; response mechanisms for abnormal gas fee scenarios, with some protocols starting to design emergency response processes for abnormal network congestion, such as temporarily raising the liquidation bonus incentive to attract more liquidators willing to participate in bidding even in a high-gas-fee environment; price protection and minimum execution price mechanisms, with some protocols designing protective clauses similar to 'this batch of collateral's execution price can't fall below a certain proportion of a reference market price,' avoiding an extreme near-zero execution outcome; and the insurance fund mechanism covered in an earlier article, serving as a last line of defense even if the auction mechanism still fails.
The core logic across these improvements all points in the same direction: no longer assuming the market will always have sufficient competing participants, and instead designing corresponding defense mechanisms specifically for the extreme scenario of 'competition failure' itself.
How should an everyday depositor judge whether a lending protocol is prepared for a Black Thursday-type incident?
A few concrete indicators worth verifying: checking the protocol's official documentation or post-mortem reports for whether they explicitly mention concrete improvement measures addressing the auction mechanism and abnormal gas-fee scenarios — if a protocol that's existed for a long time has never publicly discussed how to handle a stress-test scenario like Black Thursday, that itself is worth raising alertness over; checking whether the protocol has an insurance fund as covered in an earlier article, and whether this fund's scale relative to the protocol's overall risk exposure is reasonable; checking the details of the protocol's liquidation auction mechanism design, specifically understanding whether the auction process concludes instantly or is designed to let multiple liquidators keep bidding over a period — the latter usually has greater resistance to a competition-failure scenario.
For an everyday depositor, Black Thursday's most practical reminder is: even if you yourself were entirely uninvolved in any borrowing operation, the protocol's overall liquidation mechanism design quality can still affect your deposit safety through bad debt sharing — this isn't a risk that automatically gets ruled out by 'it's not my problem, since I didn't borrow anything,' worth spending time to verify clearly before deciding how much capital to put into a given protocol.
An earlier article covered a liquidation cascade's mechanical principle — one liquidation's sell pressure pushes price down, triggering the next batch, cycling and amplifying. But mechanical principles can read abstractly. What happened on March 12–13, 2020, known as 'Black Thursday,' happens to offer a real case you can break down frame by frame, letting you see clearly how this mechanism actually plays out in the real world.
In mid-March 2020, the COVID-19 pandemic triggered a panic sell-off across global financial markets, sweeping through stocks, bonds, and oil, with crypto not spared either. Ether's price crashed from around $194 to around $110 within a single day, a decline exceeding 40% — this starting point is especially worth noting: what ignited the entire cascade wasn't a problem within crypto markets themselves, but an extreme shock from the external macro environment, also one of the most common causes of a cascade's 'initial trigger' stage.
A lending protocol like MakerDAO requires borrowers to maintain a minimum 150% collateralization ratio — ether's crash instantly pushed a large number of originally safe borrowed positions below this threshold, and the liquidation mechanism activated accordingly. By design, liquidators should have rushed to buy up these about-to-default positions, earning the liquidation bonus. This first domino looked entirely consistent with the mechanism's design expectation, but what happened next exposed the fragile point in this mechanism under extreme conditions.
Due to market panic, a huge number of traders simultaneously flooded into the Ethereum network trying to sell assets or adjust positions, causing gas fees to instantly spike dozens of times higher than usual. The liquidation auction mechanism originally assumed multiple liquidators competing to outbid each other, pushing the price close to fair market value, but at this moment, a large number of liquidators withdrew from competing due to excessively high gas fees, causing some liquidation auctions to have only an extremely small number of, or even just a single, liquidator bidding — in the most extreme known case, a liquidator successfully bought a batch of ether worth roughly $8 million market price for just 1 DAI (roughly $1 equivalent), because there was no one else competing against them at that moment.
This kind of scenario, 'buying collateral at a price approaching zero,' meant the protocol's actual liquidation proceeds fell far short of covering the original borrowed amount, generating bad debt. The bad debt gap accumulated across the entire incident fell roughly between $4 million and $5.67 million (different statistical timing showing slightly different accounting), and the MakerDAO community quickly convened an emergency discussion afterward, ultimately deciding to raise funds through auctioning additional MKR governance tokens to cover this gap, preventing the protocol overall from falling into a more severe state of insolvency.
Looking back at the entire incident, you'll find each domino directly caused the next — macro panic caused ETH to crash, ETH's crash triggered the first batch of liquidations, liquidation demand surging caused network congestion and gas fees to spike, spiking gas fees caused liquidators to collectively withdraw, liquidators withdrawing caused the auction mechanism to produce near-zero execution prices under lack of competition — the near-zero execution price was the direct cause of the bad debt. The root cause of the bad debt wasn't simply 'ETH fell 40%,' but that throughout this entire causal chain, the liquidation auction mechanism lost the competitive conditions it needed to function normally under network congestion.
Black Thursday is worth repeatedly studying precisely because it precisely demonstrates 'exactly where each layer of a cascade gets stuck,' rather than abstractly telling you 'a cascade is scary.' After understanding this case, you'll find the question genuinely worth raising alertness over isn't 'could ETH fall 40% again,' but 'when network congestion and gas fees spike, does my own liquidation protection mechanism (such as margin buffer or liquidation threshold setting) still function normally.' After this incident, most mature protocols made concrete improvements addressing the auction mechanism and abnormal gas-fee scenarios — when assessing whether any lending protocol is safe, it's worth specifically verifying whether it has designed corresponding defense mechanisms for this kind of extreme network congestion scenario, not just looking at whether the collateralization threshold is set high enough.