andpancakes

this idea

blows

AI kills should net around half the gold. Killing them should still give something because otherwise it becomes free harassment.

  Absolutly right, it is a big problem. AI kills should net much less gold, I would think possibly less than half. All the AI is realy good for is ocationaly grabing a flag. However I do have to mention the ones who rage quit early game probably suck near as bad as the AI. I have had an ocational game where team performance went up after somone quit, but that says a lot about how bad that person was.

The disadvantage of playing with an AI team mate is probably the biggest issue they need to work on to keep this game fun. Also some way of not penalizing a player for leaving after his partner(s) bail. I hate having to play for another 20 min in a dominate game when my partner feeds them 3 times in 5 min and then rage quits. Now I have to continue fighting this loosing battle with the "help" of the AI. If I bail  It goes on my record as a disconect too.

If you're going off to kill it in the first place, and you don't get anything, then it's free harassment, because it's time spent when you could have been doing something else.

In 2v2 against human, good human players waste your time by stalling you and retreat so he dones't get killed anyway. That's free harrasement too. You only need one member of your team to hunt down the AI so your other player can also stall the lone human player on the other team.

Since the AI just doesn't retreat like a human player do. If it is to ever be subsitute to human, killing it should net no reward. Even half reward is too much as you can just keep hunting down AI forever and still outfarm the other lone human player (since your other teamate would ideally stall his farming of creeps by harrasing him). At half that's still 800 gold per 30 seconds as well as experience, while the other lone human gains nothing.

Any award for killing a baby brain AI and netting reward in doing so is just outrages, and immoral like getting paid to outperform baby in chess while your friend stand laughing. Sucks to be the baby carer nextby.

So what? You want it to be 2v1? What's the point of that? By making the AI worth nothing you're at least giving the person on his own a chance. You make it sound like the AI charging at you and dying over and over is somehow useful to the other guy, it's not, you know as well as I do you can kill an AI player in a couple of bursts because they're terrible and they buy crap items rather than health stacking. There is no situation where I would trade a human ally for an AI, because the AI simply doesn't coordinate. I'm still not going to kill you 1+A1 vs 2 because I can't coordinate burst damage.

Personally, I kinda like it as it is....

If someone quits because of connection issues at the start, we remake.

If someone quits because they died too many times, yeah you're stuck with an AI but you were probably going to lose anyway because that's why he quit.

If someone quits because of connection issues in the middle of the game, tough luck.  It doesn't happen often enough to matter that much.

Giving the AI different awards doesn't scale all that well across difficulty levels.  Maybe on newbie servers it's the best guy there; why is it worth half the gold or whatever?  Also, what about when doubles are allowed?  I have to memorize which of the 3 erebus's was the color of the guy who just quit because he is "special" and not worth as much gold and so isn't as much a priority. 

In addition, the next patch is going to implement harsher statistical penalties for rage quitting and probably make these stats visible in the lobby and even lower the priority of connections with rage quitters in skirmish and pantheon so we probably won't be seeing as many rage quitters soon anyway.

none of your arguments hold well

1) u can remake regardless it isnt relevant

2)is still the same regardless of change

3)tough luck, but need the lucky team be awarded as well? double tough luck?

4)u cant see which one is AI? u call that a disadvantage ? cmon u are playing 2v1 or 3v2 etc, i cant take that comment serious

5)again how many people disconnect isnt relevant since the change only takes place WHEN people disconnect/quit

besides i dont know where u from or if u play only with friends,i see a guy drop at least once every 3 games

what u forget

Ive had a couple of games leading like 3lvls (lvl 10 vs lvl 7) and then i had a guy drop and the opposing team started to kill the AI like mad and that gave them such an advantage that they won the game. If i had just an AI without feeding them all that gold and exp, i would have still won those games being ahead so far.

Yeah it is.  It suggest they need not waste development and testing time on a feature that isn't needed at the start of games.

No, it's not.  What if whatever scale factor they give the AI makes it actually advantageous that the human player left?  If you have a human feeder on your team, suddenly there is an incentive to ENCOURAGE him to rage quit so you can have an AI with marked benefits.  If BOTH teams have one human feeder, the team that got their feeder to rage quit suddenly has an ADVANTAGE.  Why encourage rage quitting in this way?

Like I said, nonissue.  How many games crash/dc mid game?  This game is unstable, but it doesn't crash mid-game for me or most people it seems.  And if they mysteriously dc it's probably their network connection, not demigod... if that induces a loss oh well; blame them and don't expect the game to compensate for their shoddy connection... just don't play with them again.

Why not?  There's no I AM AN AI SO I AM NOT WORTH AS MUCH GOLD overlay indicating anything.  The point is simply that it creates an inconsistency in the game logic that is neither obvious nor intuitive.  It also creates an endless balancing problem; what if they improve the AI in the game?  Suddenly they have to rebalance the AI kill worth.  It just isn't worth the effort IMO.  Like I said, we want to play humans vs. humans.  If they rage quit (which shouldn't happen as often now) you were probably gonna lose anyway.  And the other corner cases (they dc mid-game "by accident" not happening often enough and dc's at the start of the game you can just remake) make this venture an unnecessary waste of dev time.

I do too.  Like I said, it was because they were probably gonna lose anyway.   Also like I said, because of the harsher penalties on disconnects this is a trend that should be taking a drastic downturn anyway.

Oh boy; totally overlooked the fact that sometimes the AI dies more than it should

I know the problem you are trying to solve, but this isn't a solution.  It creates unintuitive and unclear gameplay nuances, marks another area that requires development and rebalancing, doesn't fix the problem anyway, and attempts to fix a problem that is being dealt with better via 1.) future improvements to the AI and 2.) increases to penalties for disconnecting anyway. 

IMO, file this suggestion under "meh"

You're relatively new to the game.  You kill enemy erebus once at lvl 10 and it nets X gold.  You play another game and kill enemy erebus at level 10 and it nets (< X) gold.  You wonder "WTF!!?"  It's an unintuitive gameplay hack drummed up by the community because they think band-aids = good game design.

You're right; now they're going to be encouraged by their team to rage quit so at least the feeder doesn't cost them much of anything.

I'm not saying it's not a problem; I'm saying it's already being addressed five times over and we should reconsider it afterward.  Those five times over, btw, are

1-stiffer statistical penalties for rage-quitting

2-join penalties in skirmish/pantheon where rage quitters are lower priority connects

3-development of better AI

4-karma system (people who rage quit will get poor karma and no one will want to play with them

5-concede option so rage-quitting isn't your only way out of a losing proposition

6*-probably a player ban system is forthcoming too...

Because if it doesn't sufficiently solve the problem why bother with the development effort that could be spent on better things?  Seriously, this problem has been noted by the developers and they're already working on tackling it in an exhorbant number of ways.  Complain after you see how the existing solutions play out.

All of these points discourage rage quitting, which means no AI, which prevents AI-farming.

Think about it.

Not for me.  AI at beginning = restart.  AI at mid = either was gonna lose anyway or incredibly rare isolated occurrance that isn't worth worrying about.

The way I think of it is this: support isn't free.  If I were to say DEVS GIVE US ANOTHER MAP no one is going to argue NO WE DON'T WANT MAPS without considering the development cost (i.e. that new map came at the expense of something else).  Since there are already five solutions forthcoming AND as stated each of those three cases doesn't really warrant developer action significantly AND this proposed solution is another point of rebalancing implying increased QA constraints, I'd rather devs spent their time elsewhere.

Give us replays, clan options, fix pantheon/skirmish.  We can deal with this thing ourselves, at least for some time to come, and especially with the forthcoming enhancements.

Well, at least what I am saying looks good on paper then.

What is being proposed looks to me like a dirty hack even on paper.

And I don't see it getting implemented anyway......

My suggestion is, if this is really something you want, is that they should implement an AI handicap option (either kill reward, health, or an option for both).  Setting it at 200% means AI's get twice the health or gold reward or whatever, 50% means half.  This way, since it's a clearly-visible setting in the game options menu, it is exposed to players rather than being an arbitrary game nuance that must be discovered manually.  This would require interface changes and more than a few lines of code and testing (so again, not worth it atm IMO) but at least this is software development done right rather than gameplay hacks to appease whiners.

It also enables more options for skirmish games.  Can you win 3 humans vs 3 AI when they have twice the starting health?  Can you win 1 vs 3 AI when they have 50% health?  It's useful in more places and it's more intuitive.  It's software development done by someone with an ounce of foresight.

<--is a proponent of skirmish options in a skirmish-style game... see my "why can't we play 4-player FFA" thread