I'll copy paste some advice I put in a thread on the older forums.
I'll put it in terms of wonder open 5v5.
The most important part of teamgames is often not the game, particularly when you're talking about random lobbies. Due to the nature of custom games, a variety of factors influence rating such as map familiarity, slot familiarity, teammate familiarity, game size familiarity, etc. This means people that have an understanding of the people within their game often have the ability to turn a theoretically even matchup trueskill wise into a massively stacked game without ever changing the game balance itself. I could put Maverick against some 1400-1500 air player and more than likely result in the game ending at min 11. I could put Sui against (really any) 1800 in air and watch a similar thing happen.
Point is, familiarity with players (if you're host) can result in you winning the game before it begins. If you aren't host, it could prevent you from deciding to x up to a lost game, provided you would rather play no game than a bad game. But that is, of course, a personal decision to make.
Then there is the aspect of slots. If you deem yourself a competent player, you need to put yourself on a slot that has influence on the game state. Some maps can have every slot have an influence on the game state, such as selkie isle or sentons. Some do not, like wonder. If you put yourself on a side slot on wonder, you are essentially putting your faith into your teammates and their ability to turn a potential opening into a crush. If you as the side crush the enemy side at min 8, it doesn't matter much if the enemy is able to do so on the other side. If they are able to push beyond the mid player because he refused to react and just REALLY wanted to finish his t3 mex, then you lose and there is nothing you could do about it. If you die to a t3 gunship rush while your air player is busy making his 4th t2 pgen, then you lost.
Meanwhile, if you took the air slot, you could essentially react and interact with any state in the game. You could supplement a push that could potentially stalemate by gunship sniping a near dead acu. You could powersnipe the enemy air player in order to delay him and keep the t2 air stage going longer. You could do early air aggression or early defensive air to protect mid player bombers to secure a likely win on the side. Or you could just do the typical t3 air rush because the game seems to be going into that sort of game state.
Likewise, if you took mid, you could do a lot of the things mentioned above. You could rush air with tree group reclaim, bomb the side player engies, and easily secure a win for your side player assuming minimal interaction by the enemy mid player. You could lab the enemy air player, you could tml rush, build pd to support side player push, or just keep the enemy mid busy while your side player crushes. If your side player fails to crush, it does not necessarily mean game over for you because your slot is far more predisposed to tech and you likely have the ability to stop a swarm of units through PD and mass efficient t2.
How can you tell which slots are the "competent slots?" Well if you can't, probably don't put yourself on one. It tends to be slots that are free to eco, which are often air slots and more secure frontline slots. If your role in the game relies upon making infinite units at the start, there isnt gonna be much of a role for you beyond min 9 once you are made irrelevant thru ACU upgrades, PD, and air.
Just learn how to interact. You dump mass into units to defend or because you know an attack will succeed. There's also the scenario where there's nothing left to lose, but it's self explanatory. If nothing is left because the game is lost in 5 minutes, you might as well as look at the enemy base and go through a checklist of potential options and pick what has the highest chance of succeeding.
Defense is also pretty self explanatory. You look at what the enemy made and you look at what your team has. If they can stop it, don't worry about it. Either the enemy wasted mass in units or they will donate you mass in an attack. If you cannot stop it, make what will stop it or tell your team to make what will stop it. Learn the optimal mass counters to a variety of strategies, this only comes through developing game sense.
Knowing an attack will succeed is the least explanatory one but also one that requires you to develop the most understanding of the game. Sure it's easy to say that 40 strats kill an acu with no protection, but understanding it on a fine level like whether an ACU and 5 percies can kill an ML is what matters more and more as your rating increases more and more. If your attack fails, you lose the units invested into it as well as the mass invested into it, meaning you could have just turned a win into a loss for no real reason.
This logic also extends into attacks that you "know" will fail in the long term, but succeed in the short term. For example, a drop proxy base might fail to, well, win the game by killing the ACUs, but it accomplishes the goal of causing 3 players to devote their eco into land factory spam to deal with the mass spam of arty. These units are being sent to a corner where they will then be useless for another 2 minutes before they can manage to get to the front. However, if you do this drop and it is quickly countered before you manage a sizeable buildup of arty, you have now made yourself entirely irrelevant to the game as you are 3 minutes behind in development compared to your mirror and he can just build on that lead to force you even further behind. It's up to you to determine whether the risk is worth the reward by knowing the map, players, and your own capability.
This finally extends to eco, or the ability to keep yourself relevant. Even if you "are keeping up with your mirror" who has finally reached full t2 mexes while 5 other players are full t3 eco, it doesn't matter. You're irrelevant. You are there to soak up mass spots and everyone stopped considering you a player in the game 8 minutes ago. All you're good for now is buzzing around. Really, you should give your mass to someone else that can upgrade it faster rather than doing inefficient RAS boys or fabs around mexes. You need to keep track of how fast the game is progressing eco-wise, and recognize whether you have the ability to keep up or whether you are in a role that is built to be made irrelevant, like wonder side slot. If you are in that role, then perform your job and have faith in your team. If you aren't, then do what will win by looking for openings.
After the Game:
How can you find openings? Watch the replays and look for it. Really there aren't many games that, if perfectly balanced but you switched a 1.3k for a 2k, they wouldn't be able to win on their own. These could be small things like microing your first lab better or larger things like realizing you should have switched to t2 tech 5 minutes ago. Depending on your rating and your slot, the things that you should be considering "openings" differ. Less rating, less refined. Less obvious what the slot does, the less refined.
I disagree on voice making you better. FAF teamgames are nowhere near the level where voice communication is a necessity to truly succeed. You'd need to be playing like 2k+ tournament teamgames in order for voice communication to be putting you over the edge. I covered the most important parts that voice communication gives the illusion of helping you by in the pre-lobby part of this. 2 competent players on 1 team and you will be able to win basically any teamgame that isn't 2000 average rated. If you aren't, then it means both of you aren't as competent as you think you are and can easily find things you could be improving on, whether it's stopping the loss before the game starts or interacting better during the game.