It would be silly indeed to lose 20% of light for 1% gamut cover.
But some users can only reach as little as 90% of rec-709 without the filter. Those don't have a choice. They need the filter if they want to reach reference. PM/email me if you want more info on this, as it does't look like you've experienced this yourself, which would suggest it's not frequent indeed.
The *main* downside of using a custom profile without the filter for rec-709 is that you can't run a gamma autocal in a straightforward way, because there is no factory profile that targets rec-709 using the filter and the gamma autocal only works with factory profiles (unless this has changed with V11, I'll confirm when I have a chance to look at it).
I hope that JVC listens and will offer factory profiles for all combinations: BT2020 with/without the filter (we already have this), DCI-P3 with/without the filter (I think we only have with), Rec-709 with/without the filter (I think we only have without). We need this flexibility to be able to use the JVC Autocal to handle most situations in the best possible way. Otherwise it's possible but it's quite convoluted and sub-optimal. If you could use your influence to ask them to implement this in the future, it would be great.
The other thing we need is a profile off mode (CMS disabled, max brightness, to select the native, unprocessed gamut) *with* the filter enabled. They are still only offering profile off without the filter, which means that you can't get the cleanest path (best RGB separation) for a 3D LUT *and* get the widest possible native gamut, which is simply crazy. I know that
understands this, but it would be great if more people with JVC's ear (you!) could advocate for this. It would significantly improve 3D LUT calibrations in some cases.
THX is a bad idea for the reasons you mentioned, I was hoping that someone else would mention this.