For those not familiar, the code A qualifier has a 48 player in a round robin format. The 48 players are divided into 12 groups, where the top two players from each group advance to code S. When the qualifier began there were 17 zergs, 15 terrans, and 16 protoss, making the racial balance almost exactly even. At the end of it 11 zerg, 11 protoss, and only two terrans advanced. So what happened? The title is most of it, but there are other interesting observations to make.
As the title says, terran had a 19% win rate vs protoss. There were 36 games of TVP played, and protoss won 29 of them, terran 7. The another interesting note, the only terran player to beat a protoss in a best of 3 was cure vs hurricane in group F. He lost 2-0 to panic in the next round to drop down to code B, unfortunately.
In PvZ, protoss had a 45% win rate vs zerg, with 40 games played, protoss winning 18 of them and zerg 22.
In ZvT terrans had a slight deficit in that matchup as well, with a 45% win rate vs zerg, 33 games were played and terran won 15 of them.
Given the small number of games played, I don't think the difference between protoss win rate and zerg win rate amounts to much, but the 19% win rate for terran is very troubling.
Blizzard has made some changes to TvP sense the disaster in code A. The mothership core has had its vision nerfed to 11, time warp now costs increased to 100 from 75 energy, and photon overcharge lasts 50 seconds instead of 60. Terran has been given a slight buff to ghosts, who now start out with the equivalent of a free mobious reactor research, with ghosts getting 75 energy instead of 50. With the recent buffs to mech also, there is some hope that TvP might change up some .
That being said, I don't think that is quite enough. TvP is has always been a matchup where terran is forced to play aggressive in the mid game, because protoss late game (although not end game, mass ghost viking is superior) is so very nearly impossible to deal with. With the nerfs to MSC will decrease the power of some protoss timing attacks very slightly, but a 10 second reduction in photon overcharge is not going to solve the problem of terran being essentially unable to drop vs a two base protoss. The buff to ghosts is nice, but its a very minor buff all things considered, with its best application for some kind of sharp timing attack. It won't impact late game TvP, where is where a lot of problems are.
Another big problem in TvP is the variety of timing attacks available to protoss, that require very specific and sharp build deviations for terran. If a terran is facing two base blink the correct response is very different from say a straight six gate timing, or two base void ray + gateway timing, or a somewhat later two base colossus timing attack. in the early game, the threat of oracles in particular restricts terrans to a very narrow set of openers. Because of the MSC and photon overcharge, and also because detection is readily available for protoss with virtually all tech options, terran is greatly discouraged from aggressive one base and two base openers. Two rax does not work against a protoss that opens fast MSC because photon overcharge will shut it down before critical damage can be done (even against some nexus first openings). Some aggressive openers with widow mines can be effective, and it seems there is almost a timing for cloack banshees against some protoss openers, the problem is that detection is quick for protoss and, at the pro level scene in particular, a properly microed stalkers can deal with the aggression very effectively with minimal loses. On occasion hellion runbyes and drops can be effective, and there has been somewhat of a resurgence of these in proleague in particular. When you compare the effectiveness of these builds with some of the protoss options, like 10 gate + msc aggressive openers, fast oracles, warp prism + 3/4gate attacks, dts, or void ray attacks with gateway pressure, terran certainly feels quite lacking, and while the protoss openers can be effective even against a terran that prepares appropriately, a protoss that prepares well against terran aggression has a much better chance of shutting it down and getting ahead.
Another problem I want to talk about is the difference between high level TvP and low level TvP. At the gold - low masters level, TvP is even more difficult, and it has been so for quite some time. In standard bio play against protoss, ghosts control is crucial in securing terran a victory late game vs protoss. While the pros can reliably pull this off, especially against lesser skilled protoss, for a low level terran this ghost control is very difficult and requires very precise micro and control to execute. In an even TVP scenario is often true that if protoss lands one good storm or takes one good fight terran loses immediately, however terran often has to win at two or three skirmishes in a row to gain a decisive advantage in the late game. The difference in these battles is one part composition (also very sensitive for terran) but another part simultaneous splitting + emp + snipes in addition to good positioning that is very challenging for 95% of terran players on the ladder. The game must and should be balanced around the pro level scene, but in the game design phase some considerations need to be made for the average and slightly above average skilled players, so that a matchup is at least fun, if not mostly fair, at these levels of play. I don't know if this concern can be addressed in a balance patch for HOTS, but I do hope the folks at blizzard are looking into this issue when deciding how to design TVP in Legacy of the Void.