There are a couple of "common knowledge" things that people keep repeating, that I just don't think are true. As such, I am developing heretical beliefs about melee. I'm writing this down because there were some people with strong opinions about rapier melee from crossing over from heavy melee, and I am convinced that these opinions aren't correct or completely applicable to rapier.
*****
The "common knowledge" of a 2v1 is that, no matter who it is, the 2 should be able to stab the 1 quickly. And in a line fight, this is sort of true. If you can create a 2v1 on the end of a line, that's strictly good. However, in a skirmishing situation, I think that this is incorrect.
In a skirmishing situation, the 2 will probably eventually win the 2v1. However, there are a lot of tactics that the 1 can use to delay, especially if they are short-blade fighters.
- You can use sinusoidal motion to "dot the i"
- You can use timing and jukes to make people move incorrectly and lose distance
- You can threaten a diving attack past one of the outside edges of the 2
All of these things are somewhat guaranteed to work, if and only if you can threaten all of them. The last of them is why I think that rapier melee is different from heavy melee, and why I think that all of this that I'm saying is true. Because you can be "too close" for someone to get a good shot onto you, and we don't have spears.
The victory condition for the 2 in a 2v1 is to "cross the T", such that the 2 can attack the 1 at the same time. But the way most people do it involves a lot of communication, and communication takes time. Time means that the 1 can use one of the above tricks to mess up distance. And in order to attack, both of the 2 need to be in the correct place. So, if the 2 are waiting to get into the correct place before communicating their intent to attack, the 1 can indefinitely delay, by moving such that at least one of the 2 are out of position.
The solution, here, is probably for the 2 to learn how to anticipate the motion of the 1, and then attack immediately when they are going to be in position half a second from now. I'm not sure how this would work in real situations - I believe that it will just create opportunities for the 1 to juke harder. I don't believe this falls apart in top-level play, but I am open to the idea that in top-level play this is incorrect.
As such, if I'm correct, I think that "have 2 people jump 1 person" is a completely unworkable strategy. 2 people cannot force the kill to happen in the way that people keep saying they can, so long as the 1 gives ground.
All that said, forcing a 2v1 with the 1 as a skirmisher can be used to the 2's team's advantage. If the 1 is giving ground, whoever of the 2 is closer to the line-fight needs to peel off and hit the line from the side, as soon as the 1 is far enough back that they can only threaten the other of the 2. That way, the 2 are trading the 2v1 for a 1v1, but gaining better positioning in the rest of their team.
In a situation with equivalent numbers, this means you're trading worse numbers right now for a better position later. I think this is a bad tactical choice, until the melee becomes large enough that a one-person difference in a large line doesn't matter. Which happens perhaps at the point of a 6-10-person line or so? At that point, the tactical advantage of positioning outweighs the tactical advantage of numbers that you're giving up.
*****
The next point that I have is that in terms of tactical advantage, 2v1 is less than 4v3, which is less than 3v2. This is because the delaying tactics of the smaller side require coordination of movement. Adding people makes this harder, even if they have practiced together. The reason that 3v2 is better than 4v3, though, is because in a 3v2 you have 50% more people on the larger side, whereas in the 4v3 you only have 33% more people. This means you have less of an advantage, so the smaller side is more likely to be able to get a random kill and equalize numbers.
If I had to put a point value on various positions, a 1v1 or any disadvantageous position is worth 0, a 2v1 is worth 1, a 4v3 is worth 1.5, and a 3v2 is worth 2. I'm unsure of the value of a 3v1 or 4v1 - my intuition places it somewhere between 1.5 and 2.5, though it's an interesting thought exercise. Obviously, all of this assumes optimal play and players who know how to survive a 1v2 in the correct way.
It would be interesting to assign a point value to all numbers-based configurations of people up to 10v10. Perhaps this is an exercise for later.
This means that in a 5v5 melee, the best tactic should be a 1-3-1 split. And if your opponents are clever, they will also perform a 1-3-1 split in order to nullify the possible advantages you can gain from it and turn the melee into a series of 1v1 fights. Depending on the value of a 3v1, a 2-1-2 split might be better? Or a counter to the 1-3-1 split? But that's a thing that requires more data.
The reason that a 1-3-1 split is so good is that any other type of split creates the potential for the opposing team to create more advantageous numbers configurations, getting that sweet, sweet 3v2 or 4v3 action. In return, you at-worst give up a 2v1, which, as we discussed, is not a bad thing.
*****
All of the discussion thus far has been about numbers. But next, I want to talk about positioning.
One of the other complaints from people I disagreed with was that people would run up to the line and stop. While I agree that people tend to run more than they need to in melee, I think that running up to the line and stopping on either edge is a good choice because it forces your opponent's line to take a convex shape. This creates more exposure to more possible cross-shots on their line, which is good for stabbing.
The option advocated for was that you should keep going around the edge of the line, if you're going to go up to the edge at all. I disagree strongly. I think that doing that, with equal numbers, opens your line up to a zipper, while running behind their line only gives you the possibility of some Death From Behind. A DFB can be foiled in a number of ways, but the most crippling is if someone from their side who isn't involved in the zipper peels off backwards to take you.
If the opposing team is smart, and mentions when something is starting in the backfield, this is almost guaranteed to happen.
*****
One more positioning thing.
All this time I've been talking about numbers in terms of 2v1 and so forth. This completely ignores local numbers advantages, intentionally. I do this because I think that local numbers-advantages from positioning are innately less powerful than "real" numbers advantages. So having a slight wrap around the edge of their line in a 4v4 is nice, but if you have to give them a few seconds of having a 4v3 in order to gain it, I don't think it's worth it. It might be if your team is made of very strong defensive fighters, but that's another thing I need more data about. Usually, people get stabbed too fast for the positioning advantage to be worth it.
*****
Lastly, risk threshold.
In any melee less than 10v10, with equal sides, I think that optimal play turns into a bunch of one-on-one fights. Anything else has the potential to give up a numbers advantage elsewhere, for only gaining a 2v1 in your team's favor. And this is bad.
That said, turning into a bunch of one-on-one fights is bad for you if your team is less 1v1 prowess-y than the opposing team. This means that you need to engage in suboptimal tactics, to try to even things out. You need to create an advantage and press it, even if the likelihood of the advantage panning out is less than 50%. Because if the chance your team will win the many-one-on-one-fights fight is (say) 30%, then anything that gains you an advantage which has a chance of success greater than 30% is a good plan.
You are the underdog, and you need to make your opponents uncomfortable. Do it however you can.
*****
That's all of my melee heresies, for the moment. I could say more about terrain, but terrain is something which is hard to control and handled in weird, squishy ways depending on who the marshal is.
Have a good day, and good melee! I'll be happy to discuss these, but I am not super likely to change my mind without some form of semi-anecdotal proof.
