Very interesting, especially the comparison with W89. Two data points isn't enough to prove anything conclusively, but that gives us two intriguing outliers.Somewhat
http://www.twstats.com/en88/index.php?page=tribe&id=4709 - #1 tribe 65 + 110 changes from a disbanded previous tribe
http://www.twstats.com/en89/index.php?page=tribe&id=2637 - #1 tribe 368 changes, interestingly enough it has very similar if not identical settings to w93 so could be something related to that?
http://www.twstats.com/en90/index.php?page=tribe&id=157 - #1 tribe 161 changes
http://www.twstats.com/en91/index.php?page=tribe&id=60 - #1 tribe 180 changes
http://www.twstats.com/en92/index.php?page=tribe&id=182 - #1 tribe 304 changes
http://www.twstats.com/en94/index.php?page=tribe&id=3107 - #1 tribe 60 changes
http://www.twstats.com/en95/index.php?page=tribe&id=1884 - #1 tribe 31 changes, none of the top 3 have more than 100 changes
Any theories on which settings could be the 'culprit'? My experience is with much older and larger worlds. There the tribal changes stat mainly reflected tribal age, e.g. 500+ changes for a 100 member tribe after 5 years wasn't unusual.
EDIT: any chance it has something to do with the size and success of premades? E.g. both Screw and Noobs got a significant amount of their players out of recruiting in-world rather than pre-world. If other worlds have more stable premades, that would drop the stat significantly.