I've observed an extremely wide variation in the ratio of wins against Ifrit to weapon drops, as reported by my fellow forum-goers. This has led me to propose the following theory.

Drop rates for primal battles decrease as wins increase in a fixed time frame.

For example, if you have 5 wins on Ifrit Monday, 7 on Tuesday, 3 on Wednesday, it may be expected that you have roughly the same chance of obtaining a weapon or token drop on any of those days, despite the varying number of wins. In fact, spreading these 15 wins out to 3 a day would actually increase your chances of drops by a factor of 2/3.

I propose this based on the anecdotal evidence that I see players tackling primals over and over for a free day or two and complaining about few drops, whereas I, though I'm 7/7 on Ifrit with something like 80 wins, spread mine out over a long period of time, never getting more than 5 wins in a day.

I did not collect any hard data during these runs, but I have only beaten Moggle Mog three times with no drops, so I will definitely begin tracking those attempts and, if necessary, attempts against Garuda. It's also worth mentioning that this may apply to coffers and non-boss dungeon chests as well, so data there would be useful. In the interest of testing this theory sooner, I'd appreciate posts from the community with hard data, or even just anecdotes. I'll make an effort to tabulate whatever data I get, including my own, and offer my conclusion.

Thanks in advance and (Good luck!) on your primal fights!