The latest published expert testimony in the Wright court case are amazing.

Wright submitted several lists of the Bitcoin he claims to own to the court.

The first was claimed to be a strict over-estimate, submitted because wright claimed that he couldn't determine what coins were his until they were submitted by a bonded courier. This first list was generated by his employ using a list of supposed characteristics of Satoshi's mining provided by Wright  (but really plagiarized from old research by bitcoiner that linked together that may have been mined by a single large miner early in Bitcoin's life).  The code they used to generate this list had an embarrassing bug and as a result it also included a bunch of other coins matching a complicated additional pattern in addition to the pattern Wright specified. Since this list was supposed to be a conservative list, the fact that the bug added some more didn't seem to be a big deal, but it later became a smoking gun.

The other lists were provided months later and wright claimed they were generated years ago (e.g. 2011) and in delivered recently by the mysterious bonded courier.  These new lists were strict subsets of the first list: Unsurprising because he'd swore before the court that the first list was a superset-- that all the blocks he mined matched that pattern. However, the way they were subsets was extremely revealing.

In particular, these new lists included most of the coins that were only on the first list due to the bug!   One of the criteria used to generate the first list was that none of the coins had been spent, to match wright's story about his temporary lack of possession of the keys keeping him from spending them---  but they apparently looked only at the BSV blockchain to generate their lists and included some spent ones, and also some were just spent after their filings-- so as a result wright's list included coins that were spend when he swore he had no access. 

Finally, the "bonded courier" excluded enough of the first list to bring the total number of coins down to match Wright's earlier provably forged "tulip trust" documents ... but it did so mostly by leaving out coins in two big blocks of _consecutive txids_.  TXID's are sha256 hashes, so they're close to uniformly random.  The only way to get big gaps in a list of your txids is if you specifically filter out ranges by rejecting transactions.

In other words, Wright's court submitted address list which he claims is an authentic document from 2011/2012 was almost certainly generated by a process similar to:

1. In 2019, directing his employee to use a list of criteria identifying an early miner's blocks which Wright copied out of a webpage. The employee messed this task up, but their mistake makes any data derived from this list without correcting the bug extremely identifiable.

2. Taking that list removing some coins that had been spent (either on BSV or after the first list was created), but ... unable to predict coins that would be spent later, he failed to remove those.

3. With it sorted by txid, he removed two bug chunks of consecutive transactions to bring the total down to match his earlier claims.

4. Fraudulently submitting this list to the court as an authentic document written years ago when it couldn't have been because it was clearly derived from a list created by recently written buggy software.

The really fun part is that by the time Wright's employee testified before the court they knew about the bug in their initial list.  Wright's inclusion of these bug coins in the list he provided the courts much later must either be due to another instance of gross incompetence in his forgeries or because he'd already shared a bug-list-derrived-list with someone else (his lawyers or one of his victims) prior to learning about the bug.