Because the HB requires DC pass. And the t-piece has 3 legs that are effectively inputs and outputs so you would only need to combine it with a second t-splitter if you wanted to connect extra tvs. In fact Space (possibly Eliies too) does an H-splitter that serves this exact purpose. Other wise the t-splitter's 3 legs are sufficient for a 3 way setup. The H-splitters are over priced and can be easily replaced with two t-splitters connected with a short lead.
If you want to connect more than one extra tv with tv links you're going to have to use an amplified splitter connected to the fourth output on the H setup.
As for the option 1 option which you seem to think is the new improved way - it's only useful for saving a bit of cable. If you want to connect extra tvs via rf you're going to need the cables in optioon 2 anyway so you might as well just do it that way anyway and not worry about diplexers and or decoder compatibility.
HB does not require DC pass. The HB can be connected to the RF IN ports only if that is all that is required. A single t-piece is sufficient just for HB distribution from the Primary to the secondary decoders via the RF IN ports.
DC power from the decoders is only available on the RF OUT ports. Only need to use the RF OUT ports IF RF signal from each decoder needs to be distributed to other viewing points. And then DC power is only required IF you use tvLINKs and amplifiers and you want the decoders to provide power.
And as a safety measure it is not a good idea to feed DC power into the RF inputs on TVs. Quite few makes of TV do not provide any protection.
No I did not say the new method is better or worse, just offered it as an alternative. For some, cabling is a problem. That is why MC developed this as a solution especially where the 3 decoders are going to be run totally independent from each other. (No sharing of viewing environments or recorded content or any of the other services that may be available on each of the decoders).