Something no one has commented on yet is that realistically it would have been impossible to get both the 50-mile and 100K records in a single run. 100K (62.14 miles) is 25% longer than 50 miles/80K. That is a very large difference to be attempting both records simultaneously. When I first heard about that angle of the race setup, I thought "never gonna happen" unless it's a different guy targeting each.
It would be equivalent to trying to take down the 800m and 1000m records in the same race, for instance. A good example of someone who set both 800m/1000mrecords in separate races when nearly identically fit for each attempt would be Sebastian Coe. In 1981 he set the 1000m WR of 2:12.18 on July 11 just a month after his 1:41.73 800m WR on June 10. In the 1000m record he went through 800m in 1:44.56, then hung on but tied up a bit the last 50m. Almost a 3-second difference in his 800m split between the two races, which is huge at that distance.
Additionally, if you input Fordyce's and Walmsley's 50-mile times into a performance-equivalence calculator, they clock in as very similar to the current 100K record, perhaps better depending on the method. For example, using a calculator employing the Riegel method (uses a smooth mathematical curve so the results aren't statistically influenced by how often or how little an event is run), Fordyce's 1984 record of 4:50:51 comes out equivalent to 6:06:16 for 100K, faster than the current 100K record of 6:09:14. Walmsley's 4:50:08 equates to 6:05:22.
Jack Daniels' VO2max method equates Fordyce's 50-mile to a 6:02:28 100K and Walmsley's 4:50:08 to 6:01:35. Statistical methods that use current WRs as their base (Purdy or Cameron) rate Fordyce's/Walmlsey's records as 2-3 minutes slower than the current 100K record. The hitch being that WR-based methods work well for frequently run distances with a deep and long history, but not so well for distances like ultras which have much less history/depth. Below is a calculator giving results according to a few different methods:
https://tools.runnerspace.com/gprofile.php?do=title&title_id=801&mgroup_id=45577
From watching the race online and listening to interviews afterward, it was clear to everyone after about the halfway mark that temperatures would slow everyone down as the day progressed. Walmsley said he therefore adjusted mid-race to target the 50-mile record, and also for two other reasons. One was that this would be one of his best/only chances to get it, and he'd have other opportunities to attempt the 100K. (50 miles isn't officially run nearly as often as it was back in Fordyce's day.)
He also didn't want to trash himself too much for the Western States 100 in just 8 weeks, where's he's defending. He didn't mention the following, but it's going to be an abbreviated buildup as it is. He'll need at least a week of recovery first (two would be better after a 50-mile race + 20k jog), and at least several days' taper on the back end (a full week much better), so realistically he will have only about six to seven weeks to prepare. Don't envy him that at all.
Everybody suffered from the warm temperatures in this race after halfway. Look at what happened to two-time World 100K champ Yamauchi. He went through 50K in 3:00:34 ahead of WR pace, made it okay to about 48 miles, then cratered and slowed from 6:00 pace down to 6:30+ pace from there to 50 miles, and then 7:00+ pace the last loop of the course (4.7 miles), finishing 10 minutes over the WR 6:19:54. Still a good performance, but chances at a WR kiboshed. It seems to me Walmsley triaged things well — he was sweating like a dog himself the last hour and half of the race.