So far preseason changes have impacted one role considerbly

Markmen's win rates have shot up so high, followed only by assassins. Bruisers and tanks are lower than ever. With enough sample size of games now, it's fair to say that markmen especially are winning a disproportionately large amount of games. Why is that? Well if you look at the jungle changes and other changes it is pretty clear. The first dragon buff gives marksmen their power spike earlier, not later. More damage also favors the assassin class, but not as much. Mages overall are doing better than they had too but not as much as either of these roles.

Marksmen have even more edges in the preseason however, from cheaper buying options, to more gold income from last hitting, to the fact that the jungle is way harder now meaning the frequencey and effectiveness of ganks has dropped drastically which leaves marksmen more time to farm and more control of the outcome of their game.

If you look over the last 7 days on op.gg statistics page, these champions for gold+ play are consistently in the top 10 or even top 5: Ashe, Sivir, Jinx, Miss Fortune. All have no great escapes, and are marksmen. The lifting of the threat of early ganks has really really helped these champions win rates.

Win rates with high enough sample size are very indicative of a champion or even a roles game health and balance especially when taken from higher level play. I fear that Riot's response will be to nerf the champs that bubble to the top, instead of addressing actual design issues for the other roles that are currently suffering, which is a terrible solution. Shyvana for example has been bottom 20 tier for most of season 4 (except the beginning before her nerfs) and is in no better a spot now, maybe even worse.

The dragon buff is a huge problem. The damage buff affects some roles and champions way better than it suits others and its power as a percentage based damage boost will disproportionately favor damage dealing champions by a large margin. If it was flat damage, that would be different. That would actually somewhat favor tanks as they don't usually itemize for damage and extra would increase their threat. To be fair, it should give boosted damage and boosted resistance to damage. perhaps 3% to both and split it. This way it scales with most champions in ways that are beneficial so you don't have to re-balance entire roles of champions with number changes. It has to be on the first buff as well, or it will still skew win rates toward whoever gets the buff earliest.

This sort of buff has another drawback tho, it would hurt 100-0 champions who depend on their numbers to perform at all, and as we know, 3% damage doesn't cancel out 3% decreased damage (103 damage * 97% for example is 99.91 meaning you'd still do less damage so perhaps it should be balanced with that in mind).