Riot internally tests all of this stuff. They've tested champions without the stats that runes provide, both with the new system and the old system. If they buff a champion, they aren't doing it off of pure theory, they're doing it because they have an internal sample size that they can pull data from.
Obviously their internal data doesn't compare to the data that the PBE and live give, but it's enough to try and push some changes. If it ends up being to much, they can always shell out nerfs to individual champions.