Why is it that the Man Cannon boost is so inconsistent? Some of the time, they send players right where I need to send them. The rest of the time, however, the boost is either stronger or weaker, over- or under-shooting players to their deaths. Is this the cannon itself or does it have to do with variables related to the appraoch by the player? Is there anything that can be done to make sure the Man Cannon will work the way it needs to work every time?
Generally, if you approach it from the front, it'll shoot you a little shorter of a distance than if you approach it from the back. Also, the depth of the man cannon into geometry affects its strength.
The trouble I've been having is that when approaching the same [untouched] Man Cannon from [approximately] the same angle and speed, the resulting boost is inconsistent. Sometimes it works, sometimes it doesn't. And for no apparent reason.
It's because of host connection, when online man cannons affect players differently than they affect the host of the game. Try making your map offline.
I actually have been forging offline (paranoid that someone will come in and throw me off), and the problem persists. And I don't mean to keep shooting down theories, I'd just really like to solve this annoying problem.
My solution: Man Cannons are unreliable and more often than not slow down gameplay; get rid of the Man Cannons. Seriously though, I can guarantee that there are a multitude of other ways to fill the gap that you are trying to bridge without using Man Cannons or Teleporters.
Man cannons are frustratingly inconsistent and it's not truly fixable. I prefer to put mine in places that you are basically forced to approach them from one angle, because jumping from the sides or front is the most common thing that causes them to malfunction. It also helps to test them a lot with friends online and see how the host connection affects performance; what usually seems to happen if anyone has a bad connection is that the man cannon sends them a little short of where you mean for them to go, so try to adjust it to send them farther - but to not send you too far even when you're the host. The real secret weapon though is always, if possible, directing your man cannons toward the blue side of one or more one-way shields. They can make up for a lot of inaccuracy on the part of the man cannon and still put the player where you need him to go. What I'd love to see in Halo 4 is a man cannon that disregards the physics engine (unless you hit something in mid-air of course), and just sends the player to a fixed landing point. We'd only need a single man cannon object and then a corresponding hill marker labeled as the cannon's landing point; then the game would automatically adjust the stream's direction and strength to get you there, and theoretically host connection and angle of approach would have no impact on where the thing sends you.
I don't see how that would solve the problem in the long run. It seems this would only solve the problem in a very select environment. I am sure he doesn't intend for the map to be restricted to that environment. My experience tells me that the man cannon can behave differently given your approach, speed, jump(or not) onto the man cannon, phase of moon, etc. The rule I follow is give the landing pad plenty of margin for error. It has always been my understanding that the man cannon is not a sniper, but a shot gun, and the aiming is never all that great.
In the case of small maps that's true, but then the map is usually small enough that the mancannon precision is not an issue. In a game of BTB, they can speed up the game a lot though and I assume this is the type of map because it's where mancannon jumps are long enough to make accuracy an issue. I found using the strong mancannon at a lower angle in place of weaker one at a normal angle can make the jump much more consistent, reduces the jump time so it doesn't slow down the game, and eliminates reduces the problem of fall damage.