Eh, the issue is that when what dragons spawn has too much impact, and certain compositions are so much better at taking them at different phases of the game. If you get a successful bot lane gank in the early game, and it's first dragon Ocean, you've basically secured the second dragon as well, because ocean dragon means no mana/hp problems in the laning phase, so you can get easier ganks and snowball the lanes. If the next dragon is infernal, then your team becomes relatively more powerful in team fights and scales better into the late game. It makes having a strong early game even more important in solo queue, just for the chance that dragon RNG lets you snowball into the win.
The other issue, particularly with infernal drake, is that each dragon is full value, which means even more scaling. If infernal was at reduced value for each subsequent infernal drake, it wouldn't quite be as much of an issue for snowballing the game.
RNG dragons would be fine if they all had similar impact on teamfights with diminishing returns on subsequent dragons, but, as is, that's not the case. Infernal dragon is the best for the late game, while ocean is huge in the early game, and the others are just nice to have, but don't have as much impact on the outcome of the game.
Giving up an early dragon should matter, but it shouldn't matter as much as it does with an infernal or ocean drake, and the later dragons should be relatively stronger for a team that's behind to give them a better chance to come back into the game.
