It "helps" in avoiding fragmentation in the sense that it has larger clusters, but that alone doesn't help you entirely avoid fragmentation - it just makes it worse when fragmentation does occur.
Technically, larger clusters (WBFS) means that their number is smaller, thus there is a smaller risk of them getting fragmented. Practically, you get massive file chunks and after deleting a few files, you simply cannot fit those chunks in the holes left by other files. You have free space, technically, but you cannot use it. Moreover, you waste space.
The smaller the cluster size the bigger the chance that you will be able to fill up the gaps and avoid wasting disk space, but then, you have a reversed situation.
This is why there are default, recommended values - recommended by IT specialists who were in the field long before many of us were born.
Moreover, WBFS is not defragmentable via standard means as it is not visible for the OS as valid - the only thing at your disposal are custom tools with debatable benefits.
In the (likely) event of disk failure, you may as well forget about the data unless you feel like digging for it yourself manually - have fun digging in RAW. With FAT32 and NTFS, you can use standard data recovery software, as well as scanning software and all sorts of defragmenters - the sky is the limit.
The partitions are also not accessible to anything other than USB Loaders - this means no media files, no Gamecube backups, no homebrew, no nothing.
WBFS was a necessary evil - it was a temporary measure until better solutions surface when the corresponding libraries are ported and tested. They have been, which is why WBFS is no longer in development by its original creators, however *some* people cling to it as if it was their saviour.
I am still unsure whether the "benefit" of using WBFS, which would be higher read speeds, is even worth it, but that will be benchmarked soon enough to dispel any further confusion.
As of today, GBATemp's and any reputable coder/modder's stance is that it's obsolete.