The v1 site’s “Level-up in” is showing the fastest possible, which is the 23h you mentioned. Technically, it actually shows you the greater of (a) the time remaining before you reach your average level-up time, or (b) the soonest you can level up.
So, for example, if your average level-up time is 8 days and you’ve been on the level for 6 days, it will generally show level-up in 2 days. But (again) if your average is 8 days and you’ve been on the level for 9 days, and your final review for this level is in 1 day, it will show the soonest you can level up… which is in 1 day.
The v2 site’s “Level-up in” is only showing the time remaining until you reach your median level-up time, because I wasn’t done creating the new “fastest possible” calculation when I created the new v2 dashboard. The v2 “fastest possible” calculation is done now, and actually takes every individual item’s current status into account, which the v1 site couldn’t do since that info wasn’t available in the WK APIv1. But I haven’t had time to go back in and update the v2 site’s “Level-up in” calculation to include the “fastest possible”.
As for the “Typical Level-up”, v2 uses your median time, and v1 uses your average time. Median is a better guess in almost all cases, because it is unaffected by very long level times (like if you stopped WK for a while). Usually, people’s median time is shorter than their average time, but there are a few exceptions.
Anyway… I’ll finish the v2 stuff eventually, but I won’t have time for at least a few more months. Other than that, the v2 site is generally much more accurate.