DiaH20
I've also been thinking of suggesting modifications to other Netherlands Antilles Gulden pages, based off data from NGC.
Given how they are one of the go-to's for coin grading, I'd be tempted to believe their stats. 🤔
Also, in regards to this, NGC's information is just taken from the Krause catalogs, which are occasionally inaccurate, especially on modern base metal coins like these where most collectors aren't interested in the finer details.
I looked into the 50 cent dates a little more, of the three dates in question, I think 2005 and 08 could have been issued to circulation, but I'm very doubtful 2007 was. If you look on ucoin, the 2007 50 Cent only has a single, set coin pictured, compared to the 1 Gulden of the same year which has several examples, including circulated ones, despite having the same listed mintage.
I'm inclined to believe that the mintages during this period just aren't 100% accurate. If you look at the 25 cent mintages relative to Numista frequencies, 2005 and 2006 have mintages over 100,000, but are both just over 1%, on par with the set-only dates earlier in the decade. 2007, on the other hand, is listed with a mintage of 2,000 and set only, but has a frequency of 13%, matching coins of other dates with around 1 million minted. Some of the 1 Gulden dates from these years are definitely more common than their listed mintages suggest, as well, most notably 2004 and 2006.