But I suppose the only really reliable thing are the statistics.
When the current modified version came out, it has only been picked a handful of times. When Pug was popular, Lavafort original was picked on a fair few occasions. It stopped being picked after the modified version. To offer a reasonable comparison, the original Asthenosphere got picked in the league but rarely on Pug. Che's modified version made the map way more popular of Pug, so I think we can use these statistics to gain a decent idea.
We had a lovely little look on Discord about the actual stats, and pulled up an archived version of league map picks that covered over 200 league matches at 3055 maps over a 5-month period sometime around 2008 or afterwards, and the old UTApug stats (I believe they ended around 2013?) which covered 12030 maps played. There is variation here between maps picked which is how the league saved their data, and maps
played which is how the PUGstats display their data. This means that many maps may have been picked and never played in PUG, so league is a better source of map popularity.
Another issue here is that PUG stats have a massive consistency in which maps got played. 8 maps accounted for nearly 92% of all maps played, and if you include AsthenosphereSE it's an astounding 96%. There was definitely more variation in league. The league's 12 most popular maps only got picked 85% of the time. The interesting thing here though is that within the league you expect maps like Oceanfloor/AL, Guardia/AL, Asthenosphere/AL, Overlord, and HiSpeed, not to get picked. This largely is true, as 7 of those maps were picked less than 1% of all maps, effectively limiting the map pool to 16 maps from 24. On PUG, 12 of 25 maps get played less than 1% of the time, nearly one half the maps. If you include the 3 maps that get picked less than 1.5% of the time then you're looking at 60% of the maps on PUG list that don't get played frequently. Again, this data is old and it's likely thanks to NoMercy's captaining DustBowl gets more attention now than it did (or should) but the fact is that the PUG is a curated maplist and the UTA maplist had antiquated map versions forced in and kept for legacy yet still maintained a greater variation of maps chosen for play.
That said, the issue of LavafortPV vs. Lavafort is interesting. The original Lavafort got picked 1.9% of all maps, barely ahead of the much-disliked RiverbedAL, and I've calculated the likelihood of it being picked to be 13.3% (pick likelihoods are higher in league because each team got 7 picks). It's difficult to draw a
direct comparison between league and PUG, but RivSE and LavaPV were nominally popular and in the top 75% of the maps on the PUG maplist with much better competition, and Lavafort and RivAL were in the top 62% of a maplist where 33% of the maps were garbage. It's not much, but if you see the below image and linked spreadsheet you can see that Lavafort and RivAL were clearly the bottom of the barrel of playable maps and both fare better-ish in the more-curated PUG list that, again, does not include all maps
picked.
Maps Spreadsheet:
https://docs.google.com/spreadsheets/d/1jpXlCD1_QGPZYKG_qanZ71zpaXJK4xt1u5wkfxvoFwc/edit?usp=sharing
Also, I've got here a map pack of some of the maps I worked on. They include a PUG-playable version of Guardia we enjoyed on #uscpug (might be fun to test IMO), the re-worked html approved™ BioAssault, the re-worked/fixed Hellraiser which is an amazing and underrated map, and a version of Saqqara with the ending fixed. All had pretty decent success on #uscpug and we played pretty often with a lot of the regular UTApug crowd.