Suicidal kills: The dos and don'ts.

The main problem is that, while you can know you're going to die if you go in after someone, you don't know you're gonna kill them. A heal, shield, or invul, a ranged attack on you from the mist, even just a character with more hp or hp regen than you thought, and you've just managed to kill yourself period. Even if you get a 90% sucess rate, the extra gp/xp you get compared to farming won't offset the massive loss if you feed yourself to an opposing player.

12 disconnects. You've played 33 games with 12 disconnects. that's over 33% to disconnect.

And? Bad internet connection, plus leaving whenever my teammates (or myself, in my first few games) fed a ton. I also dropped out a bit when the game was already lost, IE they had giants and we didn't even have the warscore for Catas. Yay. I'm sorry that not everybody has a perfect connection or has enough free time to figure that, if they have time for one game of DG, they should make it the one where they have teammates who are AFK. I also politely DC if people ask me too because my internet is lagging. So... yeah, keep making assumptions. This post is already sounding assholish.

That means. If i hosted a game and checked you out before the game. You would then be labeled a, "Ragequitter." We'd still play but i would have told you that you and I won't ever play again if you rage quit. That is probably what you do after one of your suicides failed. I have done what you have stated in this post. It's typically a tunnel vision mentality when it happens and i'm not really proud of them.

Some assumptions from the asshole post! Yay! As I've stated, it is my internet connection, or quitting when I've obviously lost. I'll freely admit I ragequit two or three times when I was just starting the game, but those were matches on when I was playing Regulus and thought that beating hard mode and having a swift anklet meant I was hot shit. Also, what are you talking about, my suicides? I just wrote a guide on when to not make suicide runs that basically said "if you have ANY doubts, don't do it, and if you even think you saw another enemy nearby, don't do it." You are acting like somehow I'm going against my own guide just so you can make a point about stats that are meaningless because of the horrible stat tracking.

The big difference between you and me...well I'm 76-22(stats don't lie). I dont usually do what you are preaching. I have the skills to follow you,kill you and get out without dying.

Dude, first off, the stats do lie. In case you haven't noticed, stat tracking is terrible. Secondly, even if the stat tracking was working, your win/loss is dependant on your team as well as yourself. If you have a horrible team of randoms, then you are going to lose. That is what happened to me on a good deal of my losses. Granted, I'm not the best player, but so what? That brings me to point 3: The math. The math works out that killing is ALWAYS better than not killing, even if you assume you could have kept the lane yourself. That's also with the assumption that you still can keep it after being battered up from your fight while the enemy DG runs in at full health. Finally, how can you have the "skills" to do that? The entire situation I presented was if YOU WOULD DIE. It is not "OK, you are going to live and the enemy is weak, do you chase him for the kill?" The situation was "when should I go into the towers to get a kill if I am so weak that I am going to die myself getting the kill?" In that situation, you won't be strong enough to hold a lane, and you can't make the towers magically not shoot you with your skill. I posted a very specific situation, and you ignored it to talk trash. Nice going.

I guess I did sound like an asshole. Mission accomplished in my book.

[

Yes, you did. It's a pity you didn't make a point and sound like somebody from the Steam forums, where you need to have 1000 hours as a certain class to even get recognition that you are worth more than the bacteria on your keyboard. Stats don't matter; they are tracked badly, and you rely on your allies as much as your enemies. In my book, as long as you can manage a positive K/D ratio in PUGs, it means you are probably better than the average player.

Oh, by the way, I'm going to have fun satirizing you now.

Your opinion is worthless because you have less karma than me. You've made very few posts, and probably came here just to troll me out of jealousy for all my great karma. I know you want karma and to sound smart like me, and even though it's meaningless, it still gives me the right to act like my opinion trumps hard facts. Also, you probably just get your wins by playing with premades in noob only PUGs on crucible while you laugh at how bad they are and ruin the community.

See how stupid that was? You probably don't, but I hope other posters do.

The main problem is that, while you can know you're going to die if you go in after someone, you don't know you're gonna kill them. A heal, shield, or invul, a ranged attack on you from the mist, even just a character with more hp or hp regen than you thought, and you've just managed to kill yourself period. Even if you get a 90% sucess rate, the extra gp/xp you get compared to farming won't offset the massive loss if you feed yourself to an opposing player.

Err... As long as you have a 75% success rate, if you are going against the DG you are chasing and another enemy who happens to be there, you are breaking even (assuming equal levels and gold being of equal importance to all players involved.) That also assumes that you don't get hit by the tower, in which case you would feed even less EXP and gold (if you got killed by the tower for two assists every time, you would have to get 50% to break even.)

That doesn't include the time it takes to get back, I admit, but on pure gold alone it's only 50% success against one DG if he can always last hit you without that, so you certainly don't need a 90% success rate when chasing; an 80% success rate is good enough that you will probably always come out on top with everything else included.

And I'm not saying you should always chase. I'm just saying, there are situations where, if you can cleanly beat the enemy DG and wind up dying while trying to LEAVE the towers afterwards, you should take it into consideration and not just obey the mantra "never die" without understanding why "never die" is considered a good thing.

Its easy to inflate stats so good ones don't necessarily indicate ability. However, given the ease of acquiring good stats a lackthereof is telling.

This is will make me sound like a snob, but in a typical public game I am usually 100% convinced that I am more effective in the field than any enemy DG, and therefore would rarely trade my life for anyone else's. In a high level game its likely that players and DGs will be relatively evenly matched, and in those situations trades aren't so bad, but they become incredibly risky. Good players are likely to have a potion, stun, ambush, Sedna rescue, or some other nasty trick to turn your suicide kill into a suicide.

Quoting SoFFacet,

reply 4

Its easy to inflate stats so good ones don't necessarily indicate ability. However, given the ease of acquiring good stats a lackthereof is telling.

Wait, since when was there an "ease" of aquiring good stats? I remember in my first 10 games or so (which weren't counted at all) I did pretty well and I think I won most of them, but none of them ever got tracked. It's just as hard to get good stats as it is to get bad ones because of the PUG format of this; you have to be absurdly good to manage to pull a bad PUG team to a victory, and you have to be extremely bad to constantly drag down even the best players (actually, you only have to be pretty bad; it's a lot easier to feed than it is to be effective.) All stats are difficult to get when stats don't track properly.

The only "easy" way I know of to get good stats is to either be so incredibly good at the game that you can drag even the worst of the worst PUGs to victory, to be so incredibly lucky you never have bad PUGs, or to play with premades. So if a player has above average stats, but not "Wow, how the hell is he so good!" stats... you can only make a general assumption that he isn't playing with premades against randoms and is somewhere between "really good and unlucky" and "average to a bit below average and pretty lucky." It's especially true considering how badly stat tracking is handled and how many weird errors there are (anybody else DCing used to make the stats weird, I believe there was a bug where if a player was dropped during startup whichever game ended first would be submitted so the player who dropped just had to compstomp to get a win, etc.)

This is will make me sound like a snob, but in a typical public game I am usually 100% convinced that I am more effective in the field than any enemy DG, and therefore would rarely trade my life for anyone else's.

Same here, honestly. I would never trade kills, because as an Ooze UB I am generally the PUG's best lane holder (unless we have a tower rook). The thing is... this isn't trading kills with an enemy. This is trading a kill on the enemy for a kill by the enemy towers. I even pointed out a lot of situations where you don't want to do it. However, to dismiss it as simply being "trading" kills is a little bit absurd; when you are getting a huge chunk of gold and a fair bit of EXP while the enemy gets nothing... I don't really see how you can call that an even trade.

If you are a good player then you will average more wins than losses in public games, but you are right that in some cases a good player can get caught on a bad team. My comment on the ease of acquiring good stats is that if you have friends that also play DG, you can assure yourself of competent teammates, skewing the stats unfairly as compared to someone of equal skill that doesn't happen to have... friends.

Still I'm amazed that DG keeps track of custom games. I can't think of many other games that count anything other than ladder.

Back to suicide kills: in a PUG, if you are the most effective DG on the field, you dying is more impactful than one of their guys dying. I know that the math says you got a better deal, but you are the one carrying your team to victory and you're gone for 30 seconds. Who knows what your teammates could let go wrong during that time?

High level games fix one problem (competent teammates) but raise another (competent opponents). As I said before, there are a lot of ways that a suicide kill can go wrong, and good opponents are the ones likely to make that happen. You always have to judge the situation, but I find it is very rarely worth the risk.

Quoting Milskidasith,

reply 3

Err... As long as you have a 75% success rate, if you are going against the DG you are chasing and another enemy who happens to be there, you are breaking even (assuming equal levels and gold being of equal importance to all players involved.) That also assumes that you don't get hit by the tower, in which case you would feed even less EXP and gold (if you got killed by the tower for two assists every time, you would have to get 50% to break even.)

That doesn't include the time it takes to get back, I admit, but on pure gold alone it's only 50% success against one DG if he can always last hit you without that, so you certainly don't need a 90% success rate when chasing; an 80% success rate is good enough that you will probably always come out on top with everything else included.

Well the two senerios are, you die and get xp/gp while your enemy dies and gets either none or some or maybe even more, or neither die but you get xp/gp and flag xp/citidel xp while your enemy does not.

The problem with the first senerio is, if you get killed instead, your enemy gets xp/gp for killing you, and his potential allies do, and he (and they) get xp/gp from farming, and xp/citidel xp from the flag, and they can team up on your remaining teammates. That's a steep price to pay to try to gain a minor advantage.

Quoting SolaceAvatar,

reply 1

The main problem is that, while you can know you're going to die if you go in after someone, you don't know you're gonna kill them. A heal, shield, or invul, a ranged attack on you from the mist, even just a character with more hp or hp regen than you thought, and you've just managed to kill yourself period. Even if you get a 90% sucess rate, the extra gp/xp you get compared to farming won't offset the massive loss if you feed yourself to an opposing player.

Not at all. If you do a quick calculation you'll find that this isnt the case. I'll run through just a quick gold estimation for early game exchanging, which is when I think exchanges happen the most.

If you successfully exchange:

You get about 1k gold. They get nothing. You lose a few waves (maybe 200 from that) that you would have gotten. Your death timeouts cancel.

Thats a net of maybe 800 gold.

If you mess it up:

They get 1k gold plus a few waves and you lose a few waves. You also lose some gold when dead.

So the net is maybe -1450 gold for you.

Even if you succeed only 75% of the time, you are coming out WAY ahead (the break even point is about succeeding 2 out of 3 times). A 90% success rate means you are making gold hand over fist.

Of course you also need to factor in assists, but I figure that the assist your team is getting from your kills about evens out the assists that they other team is getting from your deaths (and this is probably generous to the other side of the argument since assists by the guy you hopefully killed count for no gold).

Lets put it this way: if you are talking about a massive loss for feeding the opposing player in a messed up exchange, you must acknowledge the massive gain that you get from a successful exchange because its almost as large as the loss in terms of gold and XP.

And yes, of course you have to know what you are doing when you try and exchange. Like I said, its not something I recommend for new players. You certainly dont want to be trying it if QoTs or Oaks are wandering around nearby (or if you are unsure of what enemy DGs are wandering around nearby).

Back to suicide kills: in a PUG, if you are the most effective DG on the field, you dying is more impactful than one of their guys dying. I know that the math says you got a better deal, but you are the one carrying your team to victory and you're gone for 30 seconds. Who knows what your teammates could let go wrong during that time?

I suppose thats kind of true, but the bonus gold/XP you are getting more than offsets the fact the skill imbalance. Plus, like I said, on some builds the exchange is even more powerful. For example, if I'm playing as an AA Reg, gold is extremely valuable to me because once I hit Mageslayer, I get a HUGE boost up. Exchanging kills lets me get it pretty early (almost always prior to level 10) instead of a few levels later, which is a big deal.

but you are the one carrying your team to victory and you're gone for 30 seconds.

The thing is, suicide kills tend to happen early game because lategame, most towers are smashed and most DG's can handle towers pretty effectively and can run pretty effectively (at that point, it takes chainstuns to win, for the most part). At that point, people have are getting very high values of HP, armor, DPS, and mana for their money, and buying vital citadel upgrades (gold, tower upgrades, etc.).

Not only that, but the suicide kills are happening at a point where lanes are less necessary to manage; while EXP is important, 30 seconds isn't much and unlike later in the game, that doesn't guarantee you will lose a tower or have to move DGs out of their lanes to defend; everybody who isn't regulus is going to be afraid of a tower at 6k HP.

So at the point where suicide kills are most likely, I am getting the most value for my gold and I am doing it at the time where my presence is least required to keep the enemies at bay from towers/flags, and any EXP loss from losing a lane is mostly if not entirely made up for by the kill.

Edit: This is not to say you don't have a point; if you are so far beyond your teammates that they need you to help defend, then yes, keeping the lane is probably more vital. But unless your teammates can manage to really screw up in thirty seconds without you around, you probably can afford it early-game. Later in the game, suicide becomes less attractive if you are winning (if you are losing, every bit counts to either gear or creep up for a sudden push.)

Didn't know you were so sensitive Milskidasith. If the stat tracker is so bad that would mean we would all have the same shitty stat tracking which would indicate our stats are inaccurately accurate. Also...trying to be an asshole to an asshole...doesnt work. Keep on posting rather than playing.

Good luck. Have fun.

Didn't know you were so sensitive Milskidasith. If the stat tracker is so bad that would mean we would all have the same shitty stat tracking which would indicate our stats are inaccurately accurate. Also...trying to be an asshole to an asshole...doesnt work. Keep on posting rather than playing.

Sensitive? You decided to make your first post to insult me for no reason. How would you feel if I walked into your game, called you a dumbass, and left? And yes, everybody has the same shitty stat tracking. That's the thing. I have no way of knowing whether or not you are really something like 70-50 and not 70-22 (or whatever it was) like you say, but I'm giving you the benefit of the doubt. The stats are absolutely worthless.

Suicide kills are good as long as only Forces get the kill on you. Bagging and enemy kill nets you 800 gold at least. Thats as much gold as 22 waves of grunts/archers. That is a boon well worth the the 20-40s wait. The Xp you get is probably roughly what you would have gotten laning anyway.

This isn't Dota. Because the rules are different the strategies are different. OF COURSE its better to get a kill and live. But if you're sure you can get that kill before you die and not feed one of his team mates then you should go for it. Theres no death penalty but time and the rewards are vastly superior to what you would have garnered chickening out. AND you punish the guy you killed because he gets no creep waves or Xp as well, and without the benefits of a DG kill.

comparing stats is not a very useful measure.

what if i played 20 arranged matches against very high end teams and we split the matches 9 wins (my team) 11 wins (their team). i'd have a 45% win ratio and you'd call me a noob.

meanwhile you might play ridiculously easy matchups against random noobs. maybe you're even being deceptive and running your pre-made pro team against random noobs. so should i then come to believe that a player showing 150 wins and 50 losses is actually a good player?

the numbers mean shit. until i know who you played against its meaningless.

If it's a custom game stat, I don't care about it. If you can pull off good stats in pantheon/skirmish, sure, I will believe you (to some extent; bad stat tracking can be problematic). But custom games can have DCs due to bad settings, losses due to bad settings or just messing around, and wins/losses due to being against/with premades/total noobs.

Quoting Krazikarl,

reply 8

If you mess it up:

They get 1k gold plus a few waves and you lose a few waves. You also lose some gold when dead.

So the net is maybe -1450 gold for you.

Even if you succeed only 75% of the time, you are coming out WAY ahead (the break even point is about succeeding 2 out of 3 times). A 90% success rate means you are making gold hand over fist.

Of course you also need to factor in assists, but I figure that the assist your team is getting from your kills about evens out the assists that they other team is getting from your deaths (and this is probably generous to the other side of the argument since assists by the guy you hopefully killed count for no gold).

Given you're in enemy territory, I'd say on average they should be getting well more assists than you... in fact, since you're probably not chasing someone while in the red unless it's been a 1v1, your team likely is getting no assists, while kills on someone who ran into towers often gives assists to everyone on the other team.

Additionally you're losing much more than an additional 500 gp and xp. You're also losing flag time, the bonuses therein, and your allies are now outnumbered. You're basically betting 500 gp vs 500 gp, one or more flags, and possibly getting your teammates killed. That needs a very high sucess rate to be worth it. And that's ignoring assits, which are probably well in your opponent's favor, as I mentioned.

Given you're in enemy territory, I'd say on average they should be getting well more assists than you... in fact, since you're probably not chasing someone while in the red unless it's been a 1v1, your team likely is getting no assists, while kills on someone who ran into towers often gives assists to everyone on the other team.

Not at all.

You are very very rarely going to be chasing somebody who has teammates around and trying to exchange there. In order to try and exchange, you cant be around enemies who will: nuke you, stun you, slow you, heal the chasee, or shield the chasee. That means you basically arent trying to exchange with DGs around which means you arent handing out many assists. Simply being in enemy territory isnt giving anybody an assist.

On the other hand, the most common exchange situation is somebody running from a 2v1 thinking they can get back to safety but you keep going.

I find that my team probably gets about twice as many assists as the other team does on me exchanging.

Additionally you're losing much more than an additional 500 gp and xp. You're also losing flag time, the bonuses therein, and your allies are now outnumbered. You're basically betting 500 gp vs 500 gp, one or more flags, and possibly getting your teammates killed. That needs a very high sucess rate to be worth it. And that's ignoring assits, which are probably well in your opponent's favor, as I mentioned.

You are rarely losing much flag time if any at all. Like I said, usually your teammate is right behind you and the opponent is going to have to run off anyway even if you mess it up. So no flag time lost. Given that the guy has to run back and heal in any case, there isnt a very large danger for your teammates being down a player - its maybe 10 seconds or something when you factor in that the enemy has to go back and heal up almost always (or spend a pot to heal which costs them money).

Remember that if you screw up its not like the other guy is wandering around just fine. He is going to have minimal life left, which means he isnt going to be running around fighting your teammates or capping too many flags (unless he spends money on a pot to heal, in which case you've at least forced him to use up money).

Its pretty clear that the people objecting to this strategy have never tried it. They had "dont die no matter what" beaten into their head in DotA or whatever and have never even tried or seriously considered the alternatives. I've tried it both ways, and over the course of many games have come to the conclusion that I am much more valuable to my team exchanging properly. That backs up my numbers, which also unambiguously show that its worth it.

Quoting Krazikarl,

reply 16

You are very very rarely going to be chasing somebody who has teammates around and trying to exchange there.

My point being they can see you and you can't see them, so you won't know when they're going to gank you or protect their teammate.

Quoting Krazikarl,

reply 16

You are rarely losing much flag time if any at all, usually your teammate is right behind you. Given that the guy has to run back and heal in any case, there isnt a very large danger for your teammates being down a player - its maybe 10 seconds or something.

If they're already in their territory when you start chasing, it'll take maybe ten seconds for them to top off, and then your teammates have to play down a player for thirty... which means they're probably going to lose several flags, if not get killed outright.

Quoting Krazikarl,

reply 16

I've tried it both ways, and over the course of many games have come to the conclusion that I am much more valuable to my team exchanging properly. That backs up my numbers, which also unambiguously show that its worth it.

Assuming it always goes your way, it's worth it. I just think that the feeling you get from pulling off a sucessful gamble makes it seem more often than it really is... classic gambling syndrome.

My point being they can see you and you can't see them, so you won't know when they're going to gank you or protect their teammate.

It's easy to tell if there are no other enemies around; just check the rest of the map. If it's a 3v3, and two are in the other lane... then you are safe. You aren't always doubting whether or not there are enemies there.

If they're already in their territory when you start chasing, it'll take maybe ten seconds for them to top off, and then your teammates have to play down a player for thirty... which means they're probably going to lose several flags, if not get killed outright.

And if you sit back when you are at low health, it will take them ten seconds to top off and you will have to pull back anyway, or get killed when they come back. That's what the situation is.

Assuming it always goes your way, it's worth it. I just think that the feeling you get from pulling off a sucessful gamble makes it seem more often than it really is... classic gambling syndrome.

So now you say we have a gambling problem because you don't like our strategy? What the hell?

The reason events stick out in peoples mind when they win big is because the rest of it is losing small, or being average. You notice "this road always has a wreck on it" when you see a wreck two or three times, but don't notice that most of the time it is wreck free. However, in DG, both dying and getting a kill are notable events; in fact, dying is more notable, because you actually have 30 seconds to be out of the game because of it, while you have no such reinforcement that you got a kill (plus as an Erebus or UB player, dying is simply rare for me).

If you were saying we think it is more successful than not when the times it isn't successful we didn't die, I could see it. But when you die for a mistake, such a strong, memorable event, it's not going to skew the results in favor of it being successful.

risk is part of the game. being good at assessing risk can lead to advantages. sometimes the high risk high reward play is the best one by far. not sure why that's a controversial concept.

Quoting Teseer,

reply 19

If the player KILLS you after death, he will still get the gold.
(IE towers, dots, fireball mid flight and so on)

Just something to keep in mind.

like somebody already mention before this is not generally true. he will only get the money if you die BEFORE his death animation ends, meaning before its screen turns grays mode.

when he is already (lets call it) in ghost mode he won't get any mony if you die by e.g. a tick of spit or the firering.

Been thinking about this a bit more, and I think this is what it boils down to:

Going for suicide kills = generally bad.

Getting suicide kills = not so bad.

I never go after a kill that I don't think I can get away form. Sometimes I am wrong about killing the guy, but am generally right about getting away. In the heat of the moment, i think people far too often poorly evaluate the risk of dying to the rewards of getting a kill, which is why my philosophy is to always have some sort of exit plan, like "Mist form, then batswarm out in 5 seconds" or "go into the tower deadzone, and hope the guy drops a potion, or summon priests" or whatnot.

It really depends on the map. Chasing people into the towers on prison is almost guaranteed death. Chasing people into the towers on leviathan, not so much. Same goes for cataracts.. fairly large size: player maps make it more viable. Just keep an eye out for enemy portals and make sure you have a portal scroll of your own then the majority of the time you are in good shape (assuming you know what the hell you are doing to begin with).

O man, this topic is still around? I had set out a big long response but then deleted it back when this started, because I didn't think it would get legs. O well, most of it has been set out by others in the interim, but the short version (at the risk of sounding like a parrot) is:

Don't Die. Ever.

The long version is that high level gameplay excludes successful kamikaze, because a good opponent will see you gambling and punish you consistently and severely.

The best players I see will sometimes stretch the point of no return but never intentionally cross it. You don't pursue a kill if it means you die.

The most skillful games I've seen have no or very very few kills on either side until endgame, and really just come down to which team can push, farm, upgrade, and keep map control more successfully.

An AA Reg like KrazyKarl describes almost always falters in this type of game, because you just can't get the gold you need for an expensive item build until the game is over. AA/Slow Reg's primary counter - the teleport scroll

Suicide works right now because most of us are still taking wild risks and over-extending. If and when the community reaches higher levels of play it will be reduced to a Hail Mary play, a one-kill maneuver which succeeds because it is so ludicrous.

I guess since we are talking about the current state of play you could say suicide can be worthwhile if you don't value the skill of your opponents highly ^^

Another person who doesn't understand the entire concept of why don't die ever exists. If you can punish an enemy for overextending, which could easily happen, then you should, even if it means your death to the forces of whatever. Losing a kill at that point is a bad idea.

Yes, at high level play nobody will ever be in these situations. But at high level play everything is different, and even then, I specifically made sure this was written so that it only applies in situations where it should. Basically, if this situation appears, you should take it. Yes, at high level play the situation will not appear. That does not mean that at high level play, if this situation appears you shouldn't take the risk.

Hypothetical situation for you, Mil. I can't account for every single little factor but just humor me:

You are midgame in a very close high level 3v3. You have a chance to suicide kill one of their guys as he runs through his center towers on cataract. You know for a fact that neither of his buddies are close enough to save him. You do not know whether or not he has a potion. Do you go for it?

The Forums Are Now Closed!

The content will remain as a historical reference, thank you.