Bitcoin Gold Reorganizations Expose Exchange Settlement Risk

Bitcoin Gold’s transaction-finality assumptions were under scrutiny on January 26, 2020, after a technical analysis documented two deep blockchain reorganizations containing approximately 7,167 BTG in double-spends. The reorganizations occurred on January 23 and January 24, but a contemporaneous report published on January 26 brought the findings into wider cryptocurrency-market circulation.

The record supports describing the episodes as majority-hash-power, or “51%,” attacks. It does not establish that an exchange ultimately lost the full value of the double-spent coins. That distinction matters: the blockchain evidence shows conflicting transactions and replaced blocks, while the success of any off-chain withdrawal depended on an exchange’s internal controls.

Two reorganizations, 29 removed blocks

James Lovejoy’s technical report recorded the first reorganization at 18:01:32 GMT on January 23. Fourteen blocks were removed and replaced by 13 blocks, with 1,900 BTG redirected through a conflicting transaction. The report assigned an approximate contemporaneous value of $19,000 to those coins.

A second event was recorded at 00:24:08 GMT on January 24. Fifteen blocks were removed and replaced by 16 blocks. Approximately 5,267 BTG—assigned an approximate value of $53,000 in the report—was redirected through another conflicting transaction.

Together, those observations produce 29 removed blocks and approximately 7,167 BTG double-spent. The combined dollar figure of roughly $72,000 is not a venue-weighted BTG/USD closing value. It is the sum of two estimates in the contemporaneous technical report for the two attack windows, so it should be treated as an order-of-magnitude valuation rather than a verified exchange loss.

The report found that blocks in both replacement chains paid their mining rewards to the same address. That common marker supported the interpretation that the events were related. It also estimated, using NiceHash Zhash rental-price data available around the attacks, that generating each reorganization cost about 0.2 BTC, or approximately $1,700 under the report’s conversion. Those figures were estimates, not invoices or proof identifying the attacker.

Confirmations did not guarantee finality

The institutional issue was larger than Bitcoin Gold’s market capitalization. Exchanges commonly treated a specified number of confirmations as sufficient evidence that a deposit would remain in the canonical chain. Lovejoy reported that Binance credited BTG deposits for trading after six confirmations and permitted withdrawals after 12. Reorganizations spanning 14 or 15 removed blocks could therefore cross both thresholds.

The report said Binance subsequently raised its BTG withdrawal requirement to 20 confirmations. CoinDesk separately reported that the Bitcoin Gold team was communicating with exchanges and that targeted platforms had taken protective measures. Neither surviving account identified an affected exchange or documented a realized balance-sheet loss.

The episode demonstrated that a confirmation count is a risk parameter, not an absolute settlement guarantee. For a proof-of-work network with limited native hash power and compatible mining capacity available for rent, an attacker’s economic cost could be far below the face value of a deposit. Exchange controls therefore had to account for reorganization depth, deposit size, market liquidity and the cost of acquiring temporary hash power.

Later context

MIT’s Digital Currency Initiative later listed the Bitcoin Gold analysis among the outputs of its 51% Attack project and explained that its monitoring system detected reorganizations and double-spends across smaller proof-of-work networks. That institutional description strengthens attribution to the monitoring work, but it was not part of the information set available on January 26, 2020 and does not resolve whether either attempted exchange fraud produced a final loss.