Why are kill swaps considered always bad?

The problem isn't with successful kill swaps, but unsuccessful ones.  They're extremely risky to pull off, and could result in you dying and the enemy not.  Most people don't want to take this risk, or their teammates to take that risk either.

ya, Darvin is exactly right. 

if the scenario goes perfectly then you gained some free gold and you go right back to where you were before except you have extra cash.

but it doesn't always happen that way. you can botch it and die before getting the kill. you can get the kill but get picked off by an enemy DG instead of a tower so your net-gain is negated. other stuff can go wrong too, too many different ways to list them all. 

so all things being equal, if you know for certain that you can get the kill trade to go the way you want it to you should do it. but its so high risk that its hard to make a compelling case that you should ever actively try to do this, 

One of the biggest advantages of not dying is the time you get to yourself in the lane. You either have two of your team there who can really push hard on the now undefended tower/lane, or even if you just drive them back to their crystal you are still given vital moments to chip away at the towers HP - if not outright finish it.

If you die, you miss out on this really underrated advantage - and that's not even mentioning what others said above.

The risk of letting another demigod get an assist on the kill and the XP/Gold is just too much. There is nothing worse than teammates who feed the enemy. It can be incredibly hard to come back from a level deficit.

I find that typically you'll die because the enemy gets uber equipment before you die due to an experience advantage.  Experience advantages are more likely to occur because of a long-term map control scenario.  While this may be a symptom of feeding allies, it's not always the case.  Certainly it's all the harder to recover from, though.

Last game I had, I was usually 3-4 levels below my enemies for most of the match. Ended up slaughtering them. Level deficits aren't the problem. It's how you play through them. Level doesn't matter at all, it's the flags you control. Any good Erebus player should be able to tell you that.

Either way though, kill swaps have such a high risk factor involved that many times, it's just not worth it. Not only that, but it takes too much consideration of all the variables to be something worth planning around. Nine times out of ten, I'd rather let the enemy demigod escape if it means getting a flag. You can't cap flags when you're dead. As such, I'd let the enemy demigod get away.

Perfectly stated.

Yeah the only thing worse than teammates feeding is having a rage quitter on your team or someone who gives up early and quits.  So it can be pretty risky, especially if they have a regalus who might get the kill instead of the towers.

Good that I have snipe

I've been playing erebus lately so I've been going for kill swaps all the time. (\B):vulcan:(\B)

I guess I'm kinda a loose cannon for my team.

edit: post looks fine in edit mode hmmm

The other question to ask is this: Since almost all kill swaps are chasing a low hp DG into tower range, is my chance at a kill swap worth losing having the lane uncontested by that DG as he goes back to regen and effectively concedes flag control and creep gold/exp.

Problem with kill swapping is that, especially early on, it isn't actually as valuable as it feels.  If the guy you would swap with is running all the way back to base, you have two avenues:

1) Stay in lane, and get more gold and more exp due to not having a contesting DG (No need to fight, can always attack creeps, etc)

2) Try to kill the DG and Maybe get a bit of Gold.  The 250 experience you get isn't much compared to the xp of staying in the lane though.

The first option contains little to no risk.  You cap any flags there, you get gold and xp, and can potentially fight off any new DG that comes. 

Option 2 however is extremely risky.  If th enemy has a Pot and are simply being sneaky, or if you miscalculate, or any of the above, you end up losing all of option 1's benefits, and instead GIVING them to the enemy.

 truth is its not always a bad idea. 

   but for the most part its risky and a bad idea; and its not good to get in the habbit of going for gambles because sooner or later you will lose the die roll.

  Play like the Iceman, dont make mistakes, play smart and capitalize on the enemies mistakes.  In football sometimes you can pick up some extra yards by having your QB run the ball but for the most part teams dont do this because its not worth the risk of the QB getting hurt/tired ect.  Play a fundamentally sound game if its an emergency and you are far behind then maybe you need to go for some of the riskier maneuvors. 

Any strategy that involves you dying, is no such thing at all.

I play fairly aggressively as Erebus in the midgame, and get away with pursuit into or past towers more often than not.  In the early game, I don't do it.  Flags/farming are more valuable than 500 gold + level in the beginning, especially for DGs who aren't great farmers (like Erebus).

The problem is that while it feels like a good idea, can you really be sure you'll kill the enemy and die a second later? Misjudge a skill and suddenly you've died without getting a kill. At best you hurt your opponent as badly as yourself, but kamikazing stands you a very good chance to just kill yourself for no reason.