Tank bounty depending on time of last death

[Bob666]

The background:

yesterday I played some games of bt with a friend of mine who had not played it before at all. We first played vs AI, we + one AI vs 3 AI was easy. However, in another game we played 4vs4 filled with AI, and our allies had like 5-25 stats very early, I started with a troop command and wanted to change to a shredder, so I couldn't do much about it in the very beginning, and once I started into the game, the enemies already had way more gold. Still ended up with 20-10, but we lost that one.

Then, we tried to play online in a league qualify. My friend had about 0-15 stats, which he definitely didn't enjoy either. He was then kicked by his allies and I had to leave, because I didn't want to let him sit there waiting.

In the last days in general, I played some bt games again, and there was about a 40% chance, that there was a feeder in the game who decided the winner, 40% that it was an awesome game, and 20% that the teams just weren't fair from the beginning. Something like that.

So this brings me to the problem of players feeding. I thought about possibilities to solve this problems then, and came back to an idea I already had some time before, but which was never added to the game. If you don't like the idea, I would still encourage you do give any advice on how to fix that feeder problem in another way, or at least a statement why you think it cannot be fixed at all.

Feeders / bad players have to be distinguished from the the good players in the game to still let the better team win and encourage players to give their best. This is already given without any bounty because of the respawn time in which the killed player can't earn any money. We also can't punish the killed player even more, since this would make him feed even more. So feeding players need to give reduced bounty.

The solution:

My suggestion is to reset the bounty to zero, as soon as a player dies. Then the bounty grows to its original value over some given time, about 5 minutes I would say.

This means, a player dying exactly every 5 minutes won't give any more bounty in total than a player dying every 30 seconds. Stronger players who didn't die for 5 minutes will have a constant bounty from that time, so that they do not necessarily give the same amount of money to their enemies. Reaching that point may also encourage a player to be even more careful, since in the timer over 5 minutes he saves money for his team in every second. Besides that, the attackers may focus on stronger players to get more money from them, too, maybe even teaming up to prevent a player from reaching the 5 minute mark.

Also I suppose there are less deaths in games with good players, so the influence won't be too big for pro-games. In open games this could be the solution to feeding, as killing a player very frequently does not give any additional money at all. Of course, it would be easier for tinkers then and stuff, but otherwise they are hated for feeding, so this is the only possible way. As the total amount of gold gained by tank bounty will be reduced probably, increasing the base bounty could help keeping the old game length and balance. Maybe the bounty formula could be simplified then, too, since some feeder-related aspects are already taken into account, however, obviously not fixing the problem.

I know that this may change the game mechanics quite a bit and I actually doubt that this proposal gets through, but in my opinion there has to be done something about feeders. Or in general, it would be good to fix all reasons players are kicked for. And one additional note: I think it would be useless to add this as an option, since most players play on the qualify bot anyways and won't try that mode. Maybe it can be tried on the bot and removed again later, but it needs a wide start.


Thanks for reading, I hope I got some supporters by now.

[griffin1987]

make at least 2 votes from enemy team needed for a kickvote, and have people always kickable ; also give the option to leave up to 15 seconds after someone from the other team left/disconnected, without getting penalty or not getting points if your team wins (won points would just be multiplied by playedtime^2/totaltime^2)

Wouldn't that solve the problem even better?

I like thinking about the "feeder" topic, but imho you shouldn't try to minimize what feeders do to the team (which in turn would mean you wouldn't need that much teamplay anymore), but remove the possibility to kick "feeders" just because they're bad.

For me it's a really big difference if someone is feeding because he/she is bad, or because he/she suicides/doesn't care. The old kick system did a rather good job there for me and I never understood why it was changed in such a drastical way, because it worked out pretty well if there was a reasonable count of experienced players in the game - or at least, people who play for fun and want a balanced game.

So, back to your suggestion, there's one big reason I don't like your suggestion: It would make teamplay less necessary, as people would just stop to care for feeders and concentrate on their own tank way more. imho bt should encourage people to work together, to prevent deaths in team - the assist system was a very good step in that direction from my point of view.

But that's all just my .50 cent

Best Regards & Thanks for your work on the map - can still remember how I started playing nearly 5 yrs ago

[Bob666]

So, back to your suggestion, there's one big reason I don't like your suggestion: It would make teamplay less necessary, as people would just stop to care for feeders and concentrate on their own tank way more. imho bt should encourage people to work together, to prevent deaths in team - the assist system was a very good step in that direction from my point of view.

This is true, and for example I also try to take some damage for my allies to protect them (sometimes dying on my own then^^), but it is more an ideology than reality. In some cases you can do something about it, but more often the feeder brings himself in situations where you can't help him anymore.

And that is why I want the bounty to be so absolute, that just after reaching the 5-minutes-alive mark (which can of course be changed), you still have to take care of your allies to win. I don't want to change anything for those players with 5-12 stats in a team with an average of 9-9, but maybe a player with 5-17 or 5-20. Only total feed should be fixed, nothing else.

[Smell.No.Evil]

When do you kick someone?

[progg]

I love the idea Bob, but i think decreasing the bounty to 0 isn't best way. it's still a tank with hp, wepepons, able to kill, able to creep. why should it worth less than a creep. decreasing bouty to 30-40 would be better. like 15% for each death in last 5mins. i think we can discuss also increasing bounty for players who didn't die for a long time. like 10-15% for each 5mins (from base bounty couse compound interest could make bounty ridiculusly high)

---

on the other when u apply this change we should make something force players to teamplay better. maybe increase assisst bounty?

[eSVau]

Beginner Mode adjusts a players bounty by his death count, not time based but similar. And i think this mode is way tougher for those so called pro's: porter+poison+"wait for an easy target" won't help you here, you actually need to play the game

edit: well, every mode adjusts someones bounty by death count, but the impact differs much - on league mode u don't even recognize it without knowing

[AeroniumX]

I think 5 minutes is quite long. Maybe 1 or 2 minutes is ok for league? At end games you tend to die more often so I'm not sure this is a good solution. It's quite radical idea.

[UnifiedDoom]

While I admit that feeders are a problem. I think a change like this will encourage the oh-so-annoying and oh-so-arrogant style of gameplay in which players "suicide for kills." I'm sure most people have played league games in which one player has the "most kills" but also the most deaths. The player usually acts like hot stuff because he did the "most work" and demands armor upgrades and tries to kick other people because they are "useless" and "feeders" even though he has more total deaths than they do. These kinds of players are the kind that piss me off the most in btanks. To encourage such reckless gameplay I feel would be a mistake and only reward this idiotic behavior. I've always felt survivor was a more "worthy" award than "tank killer."

As for the imbalance in the games you played. I believe I was there for quite a few of them. At first I thought you were a spoofer actually but I chose to remain silent. I think you should try some ranked games before you make a judgement that "20% of the time teams were unfair from the beginning," it's not everyday a 1600 elo player randomly decides to play 5 qualify games in a row.

[Bob666]

I love the idea Bob, but i think decreasing the bounty to 0 isn't best way. it's still a tank with hp, wepepons, able to kill, able to creep.

True, but assuming the respawn time and let's say 5 seconds to engage a battle after respawn would be a minimum of about 25 seconds delay after death. Even starting at 0% of the bounty, it would have reached 25/300 = 8.3% of the at the time the tank first is in a fight. Wouldn't be much, but about 20 gold maybe in mid game. However, the starting percentage could be adjusted. While I still think, it's cleaner to start at 0% bounty, maybe raising it faster, in 3 or 4 minutes.

i think we can discuss also increasing bounty for players who didn't die for a long time. like 10-15% for each 5mins (from base bounty couse compound interest could make bounty ridiculusly high)

Well possible, but I hate surviving for some time, just to later see that I swap the lane and start to feed due to higher bounty. It's not directly associated with your suggestion, but raising bounty too high over time can lead to something like that. E.g. not dying isn't good any more, since the first death will give the enemies more money then. Of course there will be a small difference still, but it will be reduced. Regarding my idea, maybe the bounty does not need to grow linearly, it could also grow 1-0.9^time maybe, growing very fast just after the death, growing slower and slower the longer a player survives. (Though I actually like the idea of a sharp break where the bounty no longer raises at all)

on the other when u apply this change we should make something force players to teamplay better. maybe increase assisst bounty?

Yes that wouldn't be bad. Actually, as we know all attacking players, maybe the bounty can be distributed equally between all of them? Would be quite radical again of course, but this would help good assisting players to get back into the game. (As it's difficult to get a kill with a weak tank when attacking in a group)

And i think this mode is way tougher for those so called pro's: porter+poison+"wait for an easy target" won't help you here, you actually need to play the game

Yeah, not only, that weak players won't feed so much, maybe they aren't farmed so extreme any longer, because a kill isn't worth much, unless it's of strategic use.

well, every mode adjusts someones bounty by death count, but the impact differs much - on league mode u don't even recognize it without knowing

I know, but it's not changing anyones style of play or the hate against feeders I think. And my suggestion seems to be compatible with pro and noob games, rendering different modes for everyone useless. In general I dislike too many different modes, I prefer one solid mode, balanced and working for all. That's another topic, but maybe for the tank bounty a single solution for all modes is possible.

I think 5 minutes is quite long. Maybe 1 or 2 minutes is ok for league? At end games you tend to die more often so I'm not sure this is a good solution. It's quite radical idea.

Yes, maybe 5 minutes is a bit too long. It should be the border to being a feeder, maybe a bit higher than the deathrate you would kick a player for. So maybe 3 minutes, that would mean that in one hour a player could not give the enemies more money than the bounty of 20 deaths. This will probably reduce the bounty income a bit, if you look at games where you clearly won, however from the enemies' point of view those players may have been feeders, and it would still be possible to raise the base bounty a bit.

While I admit that feeders are a problem. I think a change like this will encourage the oh-so-annoying and oh-so-arrogant style of gameplay in which players "suicide for kills." I'm sure most people have played league games in which one player has the "most kills" but also the most deaths. The player usually acts like hot stuff because he did the "most work" and demands armor upgrades and tries to kick other people because they are "useless" and "feeders" even though he has more total deaths than they do. These kinds of players are the kind that piss me off the most in btanks. To encourage such reckless gameplay I feel would be a mistake and only reward this idiotic behavior. I've always felt survivor was a more "worthy" award than "tank killer."

It is true, that this allows a more aggressive style, and I like being the survivor more, too. But the reason for this is actually feeding again, since I try not to feed. If it wasn't about that gold, I wouldn't have a big problem with more aggressive players (maybe making the games shorter). Still, in case players become too aggressive, I would prefer something which doesn't damage the whole team, but the suicide player himself. Increasing the respawn time a bit for example. Currently they can die as often as they want, and the "Survivor"-player will have a hard time because of them. It's actually just another dimension of feeding.

As for the imbalance in the games you played. I believe I was there for quite a few of them. At first I thought you were a spoofer actually but I chose to remain silent. I think you should try some ranked games before you make a judgement that "20% of the time teams were unfair from the beginning," it's not everyday a 1600 elo player randomly decides to play 5 qualify games in a row.

Yes, that was me. And ranked games may be more fair. And regarding the high waiting time, only players who really want to have a good game are in there, but I don't think I am playing well enough to really do a ranked game any good. Played some ranked games already, and they were okay, but whatever, we should also try to make the qualify games good games. And if there's a good player around, it might be a challenge to fight him, but nowadays this actually leads to one player trying to kick the player with the most deaths out of his team. Just last game I had it again, an ally tried to kick another player who had 0-3 stats... I simply hate how things like that cause a fight inside a team. And the reason are feeders almost all the time. So I even think it's necessary to change something there.




So now I am defending my idea. What about other options? I think I won't support no change at all here, but other ideas are welcome. If I'm correct, these ideas are around:

- Changing bounty regarding the last time of death (linear increase, maybe above 100% after some longer time)
- Changing bounty regarding the total deaths of a player
- Increasing respawn time

- (Not influencing feeders) Increasing assist money

And I've said it before, but I would like to have a "clean" bounty formula, so that maybe the kill/death-ratio is removed from the bounty, so a player can actually know what affects his bounty.

[Smell.No.Evil]

So what about re-creating the original Death Magic?

[Velocity2k]

- Changing bounty regarding the last time of death (linear increase, maybe above 100% after some longer time)

I don't think an linear increase would be the right way to go. It should be more like a sqrt with an offset. Maybe like this:

   

[ssl]

I don't think an linear increase would be the right way to go. It should be more like a sqrt with an offset. Maybe like this:

That's exactly what I was thinking of and I think that's the best change of BTanks since a long amount of time. But still I don't like the offset. Just find a formula similar to the root that grows asymptotically to the current tank bounty. Also keep in mind while growing the bounty to the current max. bounty, the latter one changes over time!

[Velocity2k]

...But still I don't like the offset. Just find a formula similar to the root that grows asymptotically to the current tank bounty.

Well you can find more complex functions :-P

   

Code:

bounty_percent = 3.5439e-016*alive^7 - 2.9769e-013*alive^6 + 3.5722e-011*alive^5 + 2.4648e-008*alive^4 - 3.8259e-006*alive^3 - 0.0020293*alive^2 + 0.71329*alive + 25

Also keep in mind while growing the bounty to the current max. bounty, the latter one changes over time!

As this is only a factor in percent, the current bounty will always be taken into account.

[Bob666]

Just find a formula similar to the root that grows asymptotically to the current tank bounty. Also keep in mind while growing the bounty to the current max. bounty, the latter one changes over time!

I already mentioned the 1.0-0.9^time thing. I would say if the tank worth changes, we update the bounty immediataly. So just the percentage grows smoothly.

I plotted 1.0-0.99^seconds in WolframAlpha:
http://www.wolframalpha.com/input/?i=plo...3D0+to+400

In the beginning, this function is nearly linear, here is the first minute:
http://www.wolframalpha.com/input/?i=plo...%3D0+to+60

and on the long term, it will get near 1.0:
http://www.wolframalpha.com/input/?i=plo...20+to+1200

Note that it's the time from the players death, so respawning and moving to a fight takes some time too. In 1.0-0.99^seconds, the tank bounty will be at 26% after 30 seconds, which is about the complete respawn time.

Still, psychologically you would still care about feeders, because you know that dying frequently gives the enemies more money. In a linear solution, feeders just can't feed at all. However, since the function is nearly linear in the beginning, I hope that wouldn't be too bad. The most important thing is that the players notice the change, but I think since the bounty is printed to all players, it should be quite obvious. And for extreme feeders, dying very often, I would say they shouldn't give more than 40 gold each death, since otherwise it will again be a little change to the bounty, but won't change a lot about the way feeders are handled.


So am I right and this is now only about the formula, duration and offset?
What about the current modification of the tank bounty depending on the kill/death-ratio or difference, don't know. But that part which already influences the bounty despite the tank costs. I'd remove it completely.

[Prog]

2 Concerns:

1. The ability to get ahead by better play will be affected (limited).

2. Gamelength will increase.


Also there is the general worry about "caring for the worst". Why should there be anything that makes it easier for the worse player to keep up with the better player? I don't know any game or sport which does something like that. Just imagine a soccer game where goals would count less if the goalkeeper is bad and therefore there had been already 5 goals scored in the last 20 minutes. That is intuitively ridiculous. Btanks, in ranked games or tournament games that is, are in a similar fashion a competitive environment. The better way to handle it, however, is to create a non-competitive, additional environment (qualify? Novice?), where this worry does not hold.

Kicking is a different problem and I don't see why we should implement mechanics which work against the principles of competition just to get rid of unnecessary kicks. I'd rather just get rid of the kick function (again).

[Bob666]

1. The ability to get ahead by better play will be affected (limited).

This is true only for feeders. You actually get ahead if you kill your enemy every few minutes, it just won't help you to kill him even more often. In that case, you may want to change the lane and fight a real enemy or something like that. Actually better play should work throughout the game, not depending on the money you collected early.

2. Gamelength will increase.

Probably, but increasing the base bounty would change this. A 10% higher bounty should already be enough. On the other hand, players may be a bit more aggressive now (unlikely) and shorten the game. However, this can easily be fixed.

Also there is the general worry about "caring for the worst". Why should there be anything that makes it easier for the worse player to keep up with the better player? I don't know any game or sport which does something like that. Just imagine a soccer game where goals would count less if the goalkeeper is bad and therefore there had been already 5 goals scored in the last 20 minutes. That is intuitively ridiculous. Btanks, in ranked games or tournament games that is, are in a similar fashion a competitive environment. The better way to handle it, however, is to create a non-competitive, additional environment (qualify? Novice?), where this worry does not hold.

Soccer is an extremely bad example. I don't know how to project battle tanks onto soccer, but that would be comparable to a rule, that with every goal one team scores, one opponent has to leave. I mean, in Battle Tanks you build up your situation like in an RTS, unlike a shooter or soccer, where your strength just depends on your skill, not on the situation you brought yourself into. And exactly this "making strong players even stronger" isn't the best, even for tournaments. Besides that, the effect on normal games shouldn't be too big, it only changes something if a player is dying really rapidly. And I think this won't happen in a ranked game too often. I can only think of tinkers as a problem, but in fact suiciding with a tinker won't help either, since you have to pay money for the CPTP and get excluded from the game (gold and xp) for some time.

Ah, and the reason why? I think you don't know that side of the game too good, you are probably a good player and rarely in that situation where you feel like you can do nothing about the game. Battle Tanks should still be fun, and at best for all players.

Kicking is a different problem and I don't see why we should implement mechanics which work against the principles of competition just to get rid of unnecessary kicks. I'd rather just get rid of the kick function (again).

Because kicking represents the mood in which the players are in such a situation. If they were happy having a good time in that game, they would not think about kicking a player, or insulting them and so on. Kicking is just a symptom. Removing the kick-function is theoretically okay, but in rare cases players may really suicide into the enemies all the time, because they are in a fight with their team. I think it's okay to kick a player, if the whole team wants it, and maybe one enemy. Whatever, that's not the topic.


And honestly, nobody can really say what this change would bring to BT, it's a bit dangerous of course. In fact, the habit of players used to the game not wanting any greater changes made me leave BT in the first place. Progress is always a risk. And by the way, we should try to see the situation from other points of view, not only our own. It's not about making the map most fun for yourself, it is about making it most fun for all players. And noobs won't come around here, but without noobs learning the game, there won't be any new players. Don't know, I mean changing the bounty for those noobs won't help them themselves, but maybe they are accepted then and encouraged to play. I am pretty sure I would not like to learn to play BT if my first game looked like the average feeder hating in a qualify game. And there are nearly no other open BT games. And playing against bots is no option either, since only very few players really plan what they want to play online some day, they simply want it now. And don't misunderstand me, I am not trying to fix this with that bounty change, it's just the general attitude which led into this situation and which is the same now standing against the bounty change.

[Velocity2k]

Maybe we should try another approach:
Feeders often do hard on getting enough money to buy a better tank in order to stop feeding. Maybe we could decrease the gold loss a player suffers when he dies. This could be done for example by allowing a respawned player to port out of the base for free (for example for 60 seconds, cp to cp ports still cost 75 gold during this period). You still lose much money while not being able to creep.

It's only 75 gold for each port out of the base but maybe this already helps them to get a better tank faster.

[AeroniumX]

So you are going through with this idea?


If I got it right, you guys said:

Bob666: 3 minutes bounty increase to full bounty and 10% increase of base bounty.
Velocuty2k: Increasing bounty from 25% to 100% over time (according to diagram?)
Velocuty2k: Free CP teleport first 60 seconds after respawn
Maybe a miss something....


At end games high bounties allows the tide to turns faster, and adding an decrease bounty here after quick deaths will greatly prolong end games. Potentially increasing it a little. Also suicide attack enemy base tactic on the other hand might again shorten games since your bounty is lower when you recently respawned. This would be interesting anyways...

[Prog]

This is true only for feeders. You actually get ahead if you kill your enemy every few minutes, it just won't help you to kill him even more often. In that case, you may want to change the lane and fight a real enemy or something like that. Actually better play should work throughout the game, not depending on the money you collected early.

This is not only true for feeders. In fact, it is indirectly true for everyone, because the gold difference of players with different skilllevel decreases. Example: Assume the extreme case of no bounty for 3 minutes after a death. One player dies 1x/3 minutes. One player dies 2x/3 minutes. Their bounty is exactly the same. One could point to the port-/creepgold difference, but still, the difference in skill is less represented in gold than with the current system. The 1x/3 guy has, in my view, every right to be frustrated that the map doesnt recognize that he is way better than the 2x/3 player. Your proposal is less extreme, but goes into the same direction.

Ah, and the reason why? I think you don't know that side of the game too good, you are probably a good player and rarely in that situation where you feel like you can do nothing about the game. Battle Tanks should still be fun, and at best for all players.

That is a reason why there should be a place to play bt non-competitively, but that is no reason why every bt game has to have the aim to make the game fun for the bad players who do not enjoy the feeling you described. There are players who enjoy the competitive side of the game (me included). They want the gameresult to resemble the skill of all players as much as possible. And if they feed a game they try to come back somehow. There should be a place for those players as well and I think that is in the ranked/tournament/clan- games. Just like with every sport, there should be environments for those who want to compete and environments for those who just want to play around.

Because kicking represents the mood in which the players are in such a situation. If they were happy having a good time in that game, they would not think about kicking a player, or insulting them and so on. Kicking is just a symptom. Removing the kick-function is theoretically okay, but in rare cases players may really suicide into the enemies all the time, because they are in a fight with their team. I think it's okay to kick a player, if the whole team wants it, and maybe one enemy. Whatever, that's not the topic.

I think it is a really bad idea to try to cater a game to the kicking, rageing or otherwise irrational players. If you tried to remove every source that makes some players insult, kick, rage, etc there would probably be no game left. That is just how human beings are - even more so anonymous in bnet. (To give another analogy which you probably dismiss as bad as well: People rage in teamsports like soccer all the time. Would you change the rules of soccer to prevent that from happening? Is that even possible to prevent?)


Edit: Wasn't the point of the different bounty formulas of the different modes to provide some sort of "in ranked games feeder really feed" and "in beginner and normal games feeder bounty gets lower quickly"?

[Bob666]

This is not only true for feeders. In fact, it is indirectly true for everyone, because the gold difference of players with different skilllevel decreases. Example: Assume the extreme case of no bounty for 3 minutes after a death. One player dies 1x/3 minutes. One player dies 2x/3 minutes. Their bounty is exactly the same. One could point to the port-/creepgold difference, but still, the difference in skill is less represented in gold than with the current system. The 1x/3 guy has, in my view, every right to be frustrated that the map doesnt recognize that he is way better than the 2x/3 player. Your proposal is less extreme, but goes into the same direction.

Good players won't die more than once every 3 minutes, so I consider all additional deaths as "feeding deaths". It should force the good player to attack other good players on the enemy team. I.e. it doesn't pay off any longer to farm one weak player. It doesn't mean that a good player gets less money, since the bounty of good enemies will be raised accordingly.

That is a reason why there should be a place to play bt non-competitively, but that is no reason why every bt game has to have the aim to make the game fun for the bad players who do not enjoy the feeling you described. There are players who enjoy the competitive side of the game (me included). They want the gameresult to resemble the skill of all players as much as possible. And if they feed a game they try to come back somehow. There should be a place for those players as well and I think that is in the ranked/tournament/clan- games. Just like with every sport, there should be environments for those who want to compete and environments for those who just want to play around.

[...]

I think it is a really bad idea to try to cater a game to the kicking, rageing or otherwise irrational players. If you tried to remove every source that makes some players insult, kick, rage, etc there would probably be no game left. That is just how human beings are - even more so anonymous in bnet. (To give another analogy which you probably dismiss as bad as well: People rage in teamsports like soccer all the time. Would you change the rules of soccer to prevent that from happening? Is that even possible to prevent?)

You got a point there. Of course it will soften the game a bit, but I think that is something BT needs. And still, the suggestion only affects rapidly dying players, so, as said before, I think the impact on ranked games won't be too big. I don't think that it will cause any major problems, or lead to players not caring about their deaths at all. Especially the one player dying too often still needs some good playing to get back to the game, it's just that the required effort is cut at some point, where the player is truly a feeder who doesn't know what he does.

By the way, taking it to the extreme: If you gave every player a constant income, no matter of what they do, the whole game would still be skill-based, you would just have to fight players with equally strong tanks, beating them by tactics and micro. Of course this is a very extreme, this is no suggestion, but what I am trying to say is that this comparably very very little change won't be able to make a good play worthless.

Edit: Wasn't the point of the different bounty formulas of the different modes to provide some sort of "in ranked games feeder really feed" and "in beginner and normal games feeder bounty gets lower quickly"?

I know, but it's not changing anyones style of play or the hate against feeders I think. And my suggestion seems to be compatible with pro and noob games, rendering different modes for everyone useless. In general I dislike too many different modes, I prefer one solid mode, balanced and working for all. That's another topic, but maybe for the tank bounty a single solution for all modes is possible.