Long Term Thinking
Competitive battling, especially in RSE but also in RBY and GSC, is not about one factor. Winning or losing is usually dependent on three factors:
If one team has plenty of Pokémon that the other has no counters for, the first team simply has this part won. If team 1 is an Uber team and team 2 is an UU team, it looks clear who has the match. The other two parts of battling need to be in the second player's favor in order to let him have a chance.
The aforementioned situations are very extreme. Losing because of the team factor can just plain mean one player has a Dragon Dance Salamence and the other has Metagross and Tyranitar as their counters for it. We will use such an example for this guide.
This is the part you can't do anything about. The most idiotic player can beat a top-tier battler if everything goes his or her way. You can try to get more luck, but usually it's a trade-off: you trade the sure, solid Leftovers recovery for the unreliable and risky Quick Claw, or you use Thunder over Thunderbolt on Jirachi to go for 60% paralysis chance, but lose 30% accuracy. Often luck and enforcing it is frowned upon—Evasion, Quick Claw, and OHKO are the easiest examples. The more luck shows its face in battle, the more it will change the tide.
In our example, Player 2 has a Hidden Power Ice Zapdos. If Salamence gets a Dragon Dance, it can OHKO Zapdos anytime he likes. But if Rock Slide misses, Zapdos will KO Salamence. That's 'luck' not allowing the 'team' to win.
But these two examples are not what this guide will be about. There is a third part in battling that occurs in battle and where you completely have yourself to blame for if you go wrong. We call it: skill.
Skill, or the ability to play well, is, besides team building, what separates top-tier battlers from lesser battlers. I'm not going to completely discuss the term skill here: one of the biggest parts is prediction. If you want to read about that, I suggest reading skarm's General Analysis about it. Instead, I will cover another part that is a must-have: not only in RSE, but also (actually, moreover) in GSC and RBY. It's called long-term thinking. Being aware of the battle situation, and not just smashing away, keep on going to a random counter and continuing without realizing: what is happening, how's it going, what the hell am I doing? You need a plan in mind, or at least be aware whether the battle is developing to your tastes.
Salamence is up against a very vulnerable team: Exeggutor, Tyranitar, Metagross, Zapdos, Heracross, and Alakazam. As you can see, Salamence will tear this team apart as long as it is not completely foolishly played. When Salamence is brought in first time, usually after Exeggutor sleeps something, or Heracross uses Earthquake, Brick Break, or Megahorn, the opponent goes to their counter. None of his team members, however, are solid counters. In this case, the best he can do is to go to Metagross or Tyranitar. When you see someone doing that, you should make a mental note: this is his or her best Dragon Dance Salamence counter. Of this, you can conclude that it would be dumb to:
What you want to do is make a plan when Dragon Dance Salamence can sweep. It's not a good idea to Dragon Dance the first time showing Salamence anyway. You are thereby assuming he does not have a great counter, and hoping he will bring in something like Jolteon or Zapdos (stuff you cannot beat one on one without Dragon Dancing). You're much better off just firing a Hidden Power Flying or Earthquake to dent Milotic or Claydol or, in this case, Metagross/Tyranitar. Repetitive Hidden Power Flyings will force Milotic to Recover—and it's a lot easier to switch in your counter on Recover than on Surf/Ice Beam, isn't it?
Now, in this example, you should have a plan in mind. You will want to:
If you have done these two things, Salamence will have no problem killing Pokémon after Pokémon in a row. Using a sacrifice, going for the 4-0 or 5-0 is a much better and safer bet than trying to 6-0 by taking big risks that could cost you your big advantage.
But when do you have to take risks? When you're desperate and the battle is not going your way, and when this risk is going to pay off. A great, actual example is Explosion—it can take out one of your opponents, perhaps that ugly Blissey that walls your Raikou, or the Alakazam that is going to sweep you if you're not careful. But you should still be careful: an Exploder cannot counter anything anymore when it's dead. If you have Metagross/Heracross/Exeggutor/Salamence left and your opponent has an Alakazam with Psychic and Ice Punch as its only attacks, it would be very, very dumb to blow up Metagross when you need it for Alakazam. You should never move too hastily, but think twice or more before blowing up something.
In all these cases, it cannot be stressed enough that you should be aware of your opponent's choices, and how the battle is progressing. Imagine you are fighting the example team, and Sand Stream is active. You have Swampert, Snorlax, and a few relatively inconsequential Pokémon, and your opponent's Zapdos is abusing your Swampert to come in and hit you. Snorlax doesn't have Rest, and with Sand Stream it will be KOed in 3-4 Thunderbolts on the switch. While you might be at a statistical advantage (zomg i lead with 4-3 or something), the battle is not going in your favor. It is important that you can see this coming, before it's too late and Zapdos knocks off Snorlax.
Once you realize this, don't continue to play the same way, because you see it does not work. First, you know that Zapdos is important for the opponent and that it's using Swampert to come in. So, why not making it very hard to get Zapdos in? Ice Beam when you switch it in to wall Metagross or Tyranitar—they won't stay in anyway, and if Metagross Explodes on Swampert (which they have tons of counters for), Zapdos loses its easy switch-in, so you don't have much to fear. Also, since it's aiming to Thunderbolt Snorlax on the switch, it won't attempt to Substitute (especially not in Sand Stream) and it won't Hidden Power Grass if they know you will switch.
So, just Ice Beam it when they don't expect you to leave it in. Without Rest or Wish support it won't stand this, and you will play a 50/50 mind game instead of having your last counter toasted. Now, Zapdos will have to be doubly careful about switching in, and about using Thunderbolt. Snorlax has a MUCH easier time coming in on a 70 Base Power un-STAB Hidden Power Grass than on a 95 Base Power STAB Thunderbolt. If Zapdos doesn't have a way of recovery, you have a chance to completely beat it up before it kills you. If Zapdos does have Rest, it will have a harder time coming in on anything seeing as its usual switch has a super effective Ice Beam. If they use Celebi or Blissey to heal it, abuse these turns and fire Focus Punches, set up something (like the aforementioned Dragon Dance Salamence), and make them scared about switching in. If Zapdos tries to survive a Rest, see it coming in on Earthquakes, and bring in something to abuse these free turns. Don't give this dangerous Pokémon the time to recover—and remember that a Pressured Ice Beam has 8 PP at max. Keeping track of PP is a part of long-term thinking too!
Having a view of the battle is a big advantage, so not knowing is a disadvantage. This is why it's never a bad thing to save a sweeper or something threatening for much later on. It makes an opponent less tempted to sacrifice a halfway viable Pokémon, and can make them unsure of what to do. You can save something like a Dragon Dance Salamence until you identified and weakened or fainted its counters, be them Milotic or Weezing or perhaps nothing so specific and tailored towards stopping Dragon Dance Salamence. You can only determine this, though, by saving Salamence until you've seen your opponent's entire team, which is just another benefit of thinking in the long term during your battles.
I hope the point is clear—you can never employ long-term thinking enough in a battle. Though some people might disagree with some of the battling ideas stated here, the general sentiment is shared by all solid competitive battlers.