Player Score Floor

[Metsfanmax]

Concise description:

• Introduce a score floor for all players, determined by their highest score. A player's score may not drop below this floor.

Specifics/Details:

• Initialize a floor for every player equal to 400 points less than their highest-ever score (truncated to the nearest hundred).
• A player's score may not ever drop below their floor.
• If a player stays within 100 points of their floor for 15 games, on average, the floor drops by 200 points.

How this will benefit the site and/or other comments:

• Massive point dumping becomes difficult or impossible.
• If an unlikely event makes the player unable to take their turns, and they have a lot of active games, they are protected from returning to a devastated score.
• Severe streaks of poor luck do not result in a player score that is unfair to both them and other players.

Note: all of the specific numbers used in this post are nominal and can easily be changed. Also, the depth of the floor can be based on the player's current rating.

[agentcom]

Why? Point dumping happens infrequently and is usually picked up on pretty quickly. The dumper and the dumpees are punished swiftly.

And how often does someone have to disappear from CC long enough for it to ruin their score? I'm guessing that this, too is a rare occurrence.

However what you would be allowing in the back door is the "pointless" games that you like so much. As long as someone plays all their "fun/practice/whatever" games within a couple weeks, they only lose 400 points. If they've already lost a few hundred in a bad run, then the costs are even lower.

Further, someone who has gone on a bad run, maybe an assdoodler for example, can play further games with reckless abandon potentially lowering the enjoyment for others on the site.

If I had to guess, I'd say that these problems would actually occur with more frequency than the problems this was meant to solve.

[Metsfanmax]

Why? Point dumping happens infrequently and is usually picked up on pretty quickly. The dumper and the dumpees are punished swiftly.

cf. TheCrown

And how often does someone have to disappear from CC long enough for it to ruin their score? I'm guessing that this, too is a rare occurrence.

Sure, it's rare. These systems are generally designed so that it's generally very hard for you to hit your floor unless you're intending to, but that buffer is nice. It's only a secondary motivation though. The primary motivation is that people should be playing at the rankings they really deserve. It is not fair to other members of the CC community if a good player is well below the ranking that represents their skill, because they will take more points from the other players.

However what you would be allowing in the back door is the "pointless" games that you like so much. As long as someone plays all their "fun/practice/whatever" games within a couple weeks, they only lose 400 points. If they've already lost a few hundred in a bad run, then the costs are even lower.

This was not my intention -- I don't want people to be abusing the floor for the purpose of point-free games. Presumably if we did enable the unrated games, this would negate this outlet for abuse.

Further, someone who has gone on a bad run, maybe an assdoodler for example, can play further games with reckless abandon potentially lowering the enjoyment for others on the site.

That's a fair criticism. I will think about that more.

[spiesr]

Why? Point dumping happens infrequently and is usually picked up on pretty quickly. The dumper and the dumpees are punished swiftly.

cf. TheCrown

In a few ways that incident would only have been worse with a floor in place. Since theCrown's presumed goal wasn't actually to dump points, it was a side effect of losing a bunch in row. Hitting a floor would have prevented him from losing a some points, leaving his end score higher than it was. (Since he would have started from the floor when all the wins hit.
The segues into another point. Players would now have more to gain from a temporary rise in points above their "true skill level," thus encouraging similar shenanigans on a smaller scale. Once one maneuvers them self into an, otherwise unmaintainable, elevated score then they are set. If their new floor is above their previous score level then they would now be able to keep a rank higher than they otherwise would have by simply winning 50 points at a time once month to avoid the flood they now sit on dropping.

Also, I am generally opposed to anything that would allow points to be "created" by being gained & not lost.

[Metsfanmax]

Why? Point dumping happens infrequently and is usually picked up on pretty quickly. The dumper and the dumpees are punished swiftly.

cf. TheCrown

In a few ways that incident would only have been worse with a floor in place. Since theCrown's presumed goal wasn't actually to dump points, it was a side effect of losing a bunch in row. Hitting a floor would have prevented him from losing a some points, leaving his end score higher than it was. (Since he would have started from the floor when all the wins hit.

The end score that he obtained is not what is actually damaging to the rest of the community. By hitting the floor early (which might be, say, 2400 under the current proposal) he would gain a quite a bit fewer points from each individual game that he won. Suppose he was playing people averaged 2000 -- then he would have averaged maybe 25 points per game on the rise from 1000 to 3000, which required about 100 of the 333 (?) games that he eventually won. If he had instead floored out at 2400, then the people who lost 40 or more points when he was at the beginnings of his run would have been more protected against him scalping points from them.

The segues into another point. Players would now have more to gain from a temporary rise in points above their "true skill level," thus encouraging similar shenanigans on a smaller scale. Once one maneuvers them self into an, otherwise unmaintainable, elevated score then they are set. If their new floor is above their previous score level then they would now be able to keep a rank higher than they otherwise would have by simply winning 50 points at a time once month to avoid the flood they now sit on dropping.

If we trust the score system at all, we have to assume that a player's long-term average is a fair representation of their skill. If someone gets a really lucky assdoodle streak and wins 600 points, then yes for a little while they'll have a higher score. But then they'll start sinking down to where they were before. If they want the floor to protect them, they need to consistently stay 100 points above their floor for two weeks (to protect against the possibility of just gaining the 100 points once, we'll use the average of their score over the 14 days to determine whether they are floored). If they can in fact do that, then they don't even need the floor because that's where their true rank should be.

Also, I am generally opposed to anything that would allow points to be "created" by being gained & not lost.

It is a concern, but it's not inherently a bad problem. It only becomes a problem worth worrying about if noticeable point inflation begins to occur because people are flooring a lot (which means we didn't do our job well in designing the system). Also, point resets already act as a source of point deflation.

[chapcrap]

Why? Point dumping happens infrequently and is usually picked up on pretty quickly. The dumper and the dumpees are punished swiftly.

cf. TheCrown

In a few ways that incident would only have been worse with a floor in place. Since theCrown's presumed goal wasn't actually to dump points, it was a side effect of losing a bunch in row. Hitting a floor would have prevented him from losing a some points, leaving his end score higher than it was. (Since he would have started from the floor when all the wins hit.

The end score that he obtained is not what is actually damaging to the rest of the community. By hitting the floor early (which might be, say, 2400 under the current proposal) he would gain a quite a bit fewer points from each individual game that he won.

Probably not true, because all the players he was getting points from would have hit a floor too, so he they would stop losing and he could go up quicker than he did at the end.

[Metsfanmax]

Why? Point dumping happens infrequently and is usually picked up on pretty quickly. The dumper and the dumpees are punished swiftly.

cf. TheCrown

In a few ways that incident would only have been worse with a floor in place. Since theCrown's presumed goal wasn't actually to dump points, it was a side effect of losing a bunch in row. Hitting a floor would have prevented him from losing a some points, leaving his end score higher than it was. (Since he would have started from the floor when all the wins hit.

The end score that he obtained is not what is actually damaging to the rest of the community. By hitting the floor early (which might be, say, 2400 under the current proposal) he would gain a quite a bit fewer points from each individual game that he won.

Probably not true, because all the players he was getting points from would have hit a floor too, so he they would stop losing and he could go up quicker than he did at the end.

If he's averaging 2400 and they're averaging 2000, it would take at least 25 lost games for them to hit their floor of 1600. I looked at a few of the people he played and none of them played anywhere near enough games to floor (especially considering most of them won a good portion of their set).

[agentcom]

Your same arguments about how a person who gains a lot under this system will creep back down apply equally to a person under the current system who loses a lot of points. They will gradually come back up negating the reason for this suggestion and all its problems in the first place.

As for your arguments about TheCrown's situation, hitting a floor while he was on the decline and then being able to keep rising from there would have resulted in more points. I don't see any way around that. A cook who wins the same amount of games against a single opponent will still have a lower score than a general who does the same. So, TheCrown ends up with more points. Those points have to come from somewhere.

Some of those materialize out of thin air as TheCrown loses his first games and his opponents gain points, but he loses none. But some of them come from his opponents as they win fewer points while TC sits on the floor. So, net result is that TC ends with more points, and his opponents end up with an indeterminate amount of points (could be more, could be less). Perhaps that's not a big problem with your system, but it doesn't represent much, if any, of an improvement over the status quo.

So, setting that aside, you would have people who go on an unplanned vacation from the site benefiting. You would also have some point dumping prevented. Those are the positives that remain, and I'm not sure they're really that big of a deal. As you pointed out, scores tend to correct themselves, which addresses the first issue, and the second issue is quickly and adequately handled by warnings and punishments, IMO. Further, there is no prize for losing points, so there is little incentive to intentionally cause the problem that you are trying to solve.

On the other hand, your system provides an avenue for a points increase for which there is an incentive to engage in behavior that takes advantage of it. That behavior is that a person could play more risky strategies knowing that his losses are capped. Someone will figure out a way to set up games so that within a two week period, they might lose points down to the floor, but that the wins from a couple games would offset this.

Consider one variation on this. I won a Battle Royale at one point, which put me somewhere around 3500 points. My score floor would've been 3100 points, and considering that I'm on a decent run right now, I would probably currently have a score over 3100 points and well over the 2900 that I currently have.

Let's say I then went into only playing 12-player games with only Brigadiers. Every win on this type of game is worth 220 points to me and every loss costs me about 20. (In fact, while I'm a "low" general, the points break farther in my favor). Every time I can put together two wins in a row, my max score goes up and my floor goes up. Eventually, I break into General and begin the same activity there.

Keep in mind that under this hypothetical, I'm doing everything by the book. No collusion or cheating. If hypothetical me wasn't so virtuous, I could arrange some games with some people where we allow arrange consecutive victories for the players involved perpetually increasing our max scores and score floors at a more rapid pace.

Under your system, the solution for this might be to increase the span from max to floor. This would decrease the effectiveness of the system at preventing the supposed problems that you are trying to fix. All-in-all, this seems to be a bad way of solving a minor problem--a problem that maybe doesn't need to be solved in the first place.

[DoomYoshi]

Very reactionary, but very brilliant.

This is the kind of suggestion that will have a hard time selling to the public.

[Metsfanmax]

Your same arguments about how a person who gains a lot under this system will creep back down apply equally to a person under the current system who loses a lot of points. They will gradually come back up negating the reason for this suggestion and all its problems in the first place.

No, it does not negate the reason for the suggestion, because this is not just about the fairness to the person who would hit the floor. One of the primary reasons in support of the suggestion is that a person who is well below their "actual" score will be abusing their opponents by taking more points from them than they otherwise would. A person who drops well below the score should be at, will be abusing people all the way back up.

As for your arguments about TheCrown's situation, hitting a floor while he was on the decline and then being able to keep rising from there would have resulted in more points. I don't see any way around that. A cook who wins the same amount of games against a single opponent will still have a lower score than a general who does the same. So, TheCrown ends up with more points. Those points have to come from somewhere.

Some of those materialize out of thin air as TheCrown loses his first games and his opponents gain points, but he loses none. But some of them come from his opponents as they win fewer points while TC sits on the floor. So, net result is that TC ends with more points, and his opponents end up with an indeterminate amount of points (could be more, could be less). Perhaps that's not a big problem with your system, but it doesn't represent much, if any, of an improvement over the status quo.

I do not believe that TC ends with more points, though I haven't done the calculation to be certain. He probably ends up with a higher score, but I seriously doubt that he ends with more points. In other words, he went from 1000-5500 the way it actually happened, gaining 4500 points; using the floor, to top that, he would have had to go from 2400 to 6900, which is very unlikely because of the diminishing returns of the points system. So I think it's very likely that the net amount of points taken from opponents would be less, and in particular the net amount taken from any one opponent would be less, and significantly so for the first few he defeats in the run.

So, setting that aside, you would have people who go on an unplanned vacation from the site benefiting. You would also have some point dumping prevented. Those are the positives that remain, and I'm not sure they're really that big of a deal. As you pointed out, scores tend to correct themselves, which addresses the first issue, and the second issue is quickly and adequately handled by warnings and punishments, IMO. Further, there is no prize for losing points, so there is little incentive to intentionally cause the problem that you are trying to solve.

On the other hand, your system provides an avenue for a points increase for which there is an incentive to engage in behavior that takes advantage of it. That behavior is that a person could play more risky strategies knowing that his losses are capped. Someone will figure out a way to set up games so that within a two week period, they might lose points down to the floor, but that the wins from a couple games would offset this.

Consider one variation on this. I won a Battle Royale at one point, which put me somewhere around 3500 points. My score floor would've been 3100 points, and considering that I'm on a decent run right now, I would probably currently have a score over 3100 points and well over the 2900 that I currently have.

Let's say I then went into only playing 12-player games with only Brigadiers. Every win on this type of game is worth 220 points to me and every loss costs me about 20. (In fact, while I'm a "low" general, the points break farther in my favor). Every time I can put together two wins in a row, my max score goes up and my floor goes up. Eventually, I break into General and begin the same activity there.

Actually, I would set it so that it's always x hundred points below your current score, rounded down. Therefore you would need to get three wins in a row on this game type to raise your floor if you started at 3100, an improbably low occurrence. And that's assuming that the floor is as high as 400 below when you're at 3500, which is up for debate.

I grant that probably the most serious problem with this suggestion is that you can earn 100 or 200 points in a game. I'm borrowing this idea from the world of chess, where you never earn that many in a single sitting, so it may need to be modified or discarded entirely if that aspect is not resolved.

Keep in mind that under this hypothetical, I'm doing everything by the book. No collusion or cheating. If hypothetical me wasn't so virtuous, I could arrange some games with some people where we allow arrange consecutive victories for the players involved perpetually increasing our max scores and score floors at a more rapid pace.

That would be collusion, and would be against the rules, as per the TC ruling.

This is the kind of suggestion that will have a hard time selling to the public.

Evidently this is the kind of suggestion that will have a hard time selling to my own team

[agentcom]

I'd be interested to hear your findings if you actually ran the calculations and saw what happened in a variety of scenarios. I'm pretty sure that I'm right, but the burden isn't on me to prove that. I'd think you could do it pretty easily in Excel. Try out TC's system. Try out my system. Try some others, and see what happens.

Your OP says that the floor is based on your historical high score not your current score (except that it can be lowered based on your current score). That's what I was basing my post off of. I was assuming that someone does this right after they hit their high score. The chances of winning a couple games in a row during the time that they're at the floor, thus raising their max is not terribly low.

On the point about collusion, just to be clear, the first example was not collusion, simply playing certain game types with certain opponents. I intended the second part about organizing the wins and losses as an example of possible illegal collusion that would expedite the same process.

[Metsfanmax]

I'd be interested to hear your findings if you actually ran the calculations and saw what happened in a variety of scenarios. I'm pretty sure that I'm right, but the burden isn't on me to prove that. I'd think you could do it pretty easily in Excel. Try out TC's system. Try out my system. Try some others, and see what happens.

I'm not going to do a full calculation because it either involves making some assumptions about who he's playing or collecting the full record of the games played (which I'm not interested in doing). All that you need to see to get the gist here is to calculate how many games it would take him to get from where he actually started (at 1000) to the floor he would have had (2400). Assuming he's playing people ranked 2000, this would take approximately 60 games. Therefore we can take 60 of the approximately 350 he won, and assume that he would have gotten those points added to his final score of 5500. The average number of points he would have earned in this stretch is 10 per game (very roughly), so we're talking another 600 points added to his top score. Therefore as it actually happened, he gained 4500 points during the totality of his run. Had the score floor been in place, he would have earned 3700 points during the totality of his run. Score floor just saved 800 points from being taken from others. And we haven't even analyzed the number of points that are no longer artificially given to others by preventing the point dumping.

Your OP says that the floor is based on your historical high score not your current score (except that it can be lowered based on your current score). That's what I was basing my post off of. I was assuming that someone does this right after they hit their high score. The chances of winning a couple games in a row during the time that they're at the floor, thus raising their max is not terribly low

It's pretty low. Assuming generously that the chances of winning the 12 player game are 1/12, then the chances of winning two in a row are 1 in 144, and the chances of winning three in a row are 1 in 1728. More important is the fact that this needs to be done multiple times in a row to get the effect you're talking about. Suppose I start at my 3100 floor and amazingly win three in a row, raising my high score to 3760 and my floor to 3300. Now, if I want to continue raising my floor, I need to play more 12 player games and continue the impossible odds. If I play for a week or two and don't pull off the improbable again, my floor drops -- meaning I need to get the improbable luck again just to get where I started from.

On the point about collusion, just to be clear, the first example was not collusion, simply playing certain game types with certain opponents. I intended the second part about organizing the wins and losses as an example of possible illegal collusion that would expedite the same process.

I'm not particularly concerned about it because of how hard it would be (see above) and how few people would try it (since it involves spiking your own score). If this aspect of an abuse is an issue, we can rule it a game exploit to abuse the floor for the purposes of artificially inflating one's score.

I'll have to think more about how BRs play into this. It's something I hadn't considered when I wrote this up. If the floor system is to work properly, you really shouldn't be able to gain more points in one game than the distance to the floor.

[agentcom]

So under your TC analysis, he would've got 600 more points and his opponents would've got 800 more points. This makes sense because he got capped at 2400 and the additional 1400 points of losses just kind of vanish. So, I suppose I'm okay with your math there. I still disagree that this is an acceptable scenario. You've just created 1400 points that are distributed mostly to someone who has broken the rules and some more to people who helped him (perhaps unwillingly). This would encourage this type of behavior but perhaps on a smaller scale or through a different method like the one that I proposed (using multiplayer games).

But again, even if we're not looking at intentionally cheating, you could still set up your games to game the system. You are artificially lowering the odds of my argument there. First off, you could win 2/3 or 3/6. Maybe even 3/9. Still low odds, but more importantly ... your odds of winning 2 in a row aren't 1/144. They're more like 1/12. As long as you win A game while you're at your floor, all you have to do is win another one. It doesn't matter whether you lose a couple before that.

But regardless of all of that. My main objection is that this is not a big deal that needs some complex solution. How many people lose 400 points in a short period of time due to missed turns or intentionally losing games?! There are some people who have dumped points in the past. Okay, but is this really a necessary solution to correct that? There are some people who have been unable to log in to CC, but did they lose 400 points? This would require a combination of just the right circumstances: (a) they would have to be in a lot of games; and (b) they would have to miss critical turns or three consecutive turns in those games (in 8 to 12 player games, this would require weeks of absence). You can only count games that they would've won when you're calculating this "unfair" drop in points, too.

[Metsfanmax]

So under your TC analysis, he would've got 600 more points and his opponents would've got 800 more points. This makes sense because he got capped at 2400 and the additional 1400 points of losses just kind of vanish. So, I suppose I'm okay with your math there. I still disagree that this is an acceptable scenario. You've just created 1400 points that are distributed mostly to someone who has broken the rules and some more to people who helped him (perhaps unwillingly). This would encourage this type of behavior but perhaps on a smaller scale or through a different method like the one that I proposed (using multiplayer games).

This argument is misleading. Because of the way the floor works, there's not a one-to-one trade-off between points won/lost. The 1400 points he would have lost aren't a relevant number anymore, because the points given in every game are decided by the 2400 floor limit. This could give a wildly different number of total points.

But again, even if we're not looking at intentionally cheating, you could still set up your games to game the system. You are artificially lowering the odds of my argument there. First off, you could win 2/3 or 3/6. Maybe even 3/9. Still low odds,

No matter how many you are playing you still have to win two of them in a row to achieve the effect you're describing; it doesn't matter how many total you win.

but more importantly ... your odds of winning 2 in a row aren't 1/144. They're more like 1/12. As long as you win A game while you're at your floor, all you have to do is win another one. It doesn't matter whether you lose a couple before that.

In this example you need to win three games in a row to boost your floor, so that's where the 1/144 ratio becomes relevant. The odds of winning two in a row are still 1/144, it's just that you only need to make it happen once out of every 12 times to get the desired effect. And, again, you need to be able to pull this off multiple times in a row to continue boosting your score.

Your comments have made me realize, though, that number of days spent near the floor is not a useful metric for CC. It is probably significantly better to use number of games played while near the floor. For example, we could say that if you stay within 100 points of your floor after completing the first 12 games after hitting the floor, then your scenario is nigh on impossible. Even if you did win one of those 12 games, your average score would likely still be well below 3200, so your floor would drop. That might be slightly more technically complicated but it's a better solution.

But regardless of all of that. My main objection is that this is not a big deal that needs some complex solution.

And my main objection to your main objection is that this isn't a complex solution. It's pretty simple, both in theory and in practice.

[Donelladan]

If we trust the score system at all, we have to assume that a player's long-term average is a fair representation of their skill.

This is where you are wrong in the 1st place. We do not trust the score system at all.
It is totally possible to reach the top by playing only one kind of game. I think one conqueror reach the title by playing exclusively hive 1vs1.

Then the guy stop playing hive, and his score go down around major.
What would have happened with your suggestion ?
For two weeks, every major or less that would have beat him gains 50 points or more out of nowhere? He could have give tons of points away.

So your suggestion create at least two pbs-
-> You create a lot of points, therefore you increase artificially the total of points. We would have a quick inflation of points for the whole CC. For no reason.
-> You create the possibility of easy and constant abuse for a lot of people : I can lose 400 points. Then start a lot of games with my clanmate, losing all of them on purpose, and make them all win hundreds of points. And it costs me nothing because if my score is low, I'll get it high quik enough after two weeks.

It is not fair to other members of the CC community if a good player is well below the ranking that represents their skill, because they will take more points from the other players.

I really don't understand the matter here. Let's say I play a general that has gone done to lieutenant. He'll take me 30 or 40 points once, then that's fine, 30 points is nothing it's not like I am gonna play 100 games against that one and lose hundreds of points !! It's not like everyone has a constant score. We encouter a lot of fluctuations, losing a game once in a while because the guy has "a lower score that was his skill really are" is really not an important pb.
Plus, as I said as the beginning, I do not trust the score system. Anyway score system is based on all the games we play.
You can be fighting a lieutenant on a map that he actually mastered better than anyone else on CC, he still is a lieutenant and you'll lose lot on points if you are a general. On the other side, being a general doesn't mean you are a general on every map of CC. So score means not that much about skills once you go out of the map you master anyway.

[Metsfanmax]

If we trust the score system at all, we have to assume that a player's long-term average is a fair representation of their skill.

This is where you are wrong in the 1st place. We do not trust the score system at all.
It is totally possible to reach the top by playing only one kind of game. I think one conqueror reach the title by playing exclusively hive 1vs1.

Well, I'm trying to help improve the validity of the scoring system here. I believe that, if it can be implemented correctly, this will help.

Then the guy stop playing hive, and his score go down around major.
What would have happened with your suggestion ?
For two weeks, every major or less that would have beat him gains 50 points or more out of nowhere? He could have give tons of points away.

Yes, that's a concern. But that's the fault of the scoreboard itself, not the fault of the suggestion. As you say, being a general doesn't mean you're a general on every map. However, as long as we have a scoreboard where every map contributes equally, we have to work under the assumption that the scoreboard is meaningful (if it's not, why don't we just throw it out?). If you're good enough to get to Conqueror on at least one setting, we have to assume you're good enough in general to have earned the rank. (After all, even if you're playing on niche maps, getting up to 5000 points is not easy no matter how you do it.) What you're describing sounds a bit like point dumping, and it's what we're trying to avoid here. If you're OK with someone running up to Conqueror and then blowing all their points away, then I suppose you won't be convinced that this is a good idea.

So your suggestion create at least two pbs-
-> You create a lot of points, therefore you increase artificially the total of points. We would have a quick inflation of points for the whole CC. For no reason.

The total number of points is increased by 1000 every time someone joins the site; not clear to me that this is a concern.

-> You create the possibility of easy and constant abuse for a lot of people : I can lose 400 points. Then start a lot of games with my clanmate, losing all of them on purpose, and make them all win hundreds of points. And it costs me nothing because if my score is low, I'll get it high quik enough after two weeks.

That would be a rules violation, and would be dealt with accordingly.

I really don't understand the matter here. Let's say I play a general that has gone done to lieutenant. He'll take me 30 or 40 points once, then that's fine, 30 points is nothing it's not like I am gonna play 100 games against that one and lose hundreds of points !! It's not like everyone has a constant score. We encouter a lot of fluctuations, losing a game once in a while because the guy has "a lower score that was his skill really are" is really not an important pb.

Well, you differ here from many of the comments people have made in C&A. Most people oppose point dumping on this very basis -- that it's not fair to the people he plays while he's at the lower score. You may not mind, but other people do.

Plus, as I said as the beginning, I do not trust the score system. Anyway score system is based on all the games we play.
You can be fighting a lieutenant on a map that he actually mastered better than anyone else on CC, he still is a lieutenant and you'll lose lot on points if you are a general. On the other side, being a general doesn't mean you are a general on every map of CC. So score means not that much about skills once you go out of the map you master anyway.

This is a long-term problem we have to deal with. I really don't think that if you are better on a map than anyone else on CC, that you should be a lieutenant. It's just not fair to everyone else (especially if you play a lot of games on that map). But fixing that issue is much bigger than this topic.

[chapcrap]

If we trust the score system at all, we have to assume that a player's long-term average is a fair representation of their skill.

This is where you are wrong in the 1st place. We do not trust the score system at all.
It is totally possible to reach the top by playing only one kind of game. I think one conqueror reach the title by playing exclusively hive 1vs1.

Well, I'm trying to help improve the validity of the scoring system here. I believe that, if it can be implemented correctly, this will help.

Then the guy stop playing hive, and his score go down around major.
What would have happened with your suggestion ?
For two weeks, every major or less that would have beat him gains 50 points or more out of nowhere? He could have give tons of points away.

Yes, that's a concern. But that's the fault of the scoreboard itself, not the fault of the suggestion. As you say, being a general doesn't mean you're a general on every map. However, as long as we have a scoreboard where every map contributes equally, we have to work under the assumption that the scoreboard is meaningful (if it's not, why don't we just throw it out?). If you're good enough to get to Conqueror on at least one setting, we have to assume you're good enough in general to have earned the rank. (After all, even if you're playing on niche maps, getting up to 5000 points is not easy no matter how you do it.) What you're describing sounds a bit like point dumping, and it's what we're trying to avoid here. If you're OK with someone running up to Conqueror and then blowing all their points away, then I suppose you won't be convinced that this is a good idea.

No, this is really a problem with this suggestion. You'll have to get another suggestion through to change the scoreboard. Because under the current scoreboard, this one won't work.

So your suggestion create at least two pbs-
-> You create a lot of points, therefore you increase artificially the total of points. We would have a quick inflation of points for the whole CC. For no reason.

The total number of points is increased by 1000 every time someone joins the site; not clear to me that this is a concern.

That's apples and oranges. People have to start somewhere.

I'm just not really sure how this helps anything. I mean, what's the point? The only noted uses for this is to prevent the very small problem of massive point dumping and to protect people who abandon the site. Neither of these is a big benefit and this seemingly creates more problems than it solves, because of the potential abuse. I'm not sure why we're trying to fix one abuse that rarely happens to create an abuse that would happen more often. The abuse that is to be solved can already be dealt with.

[BigBallinStalin]

Pretty much agree with chapcrap's finally paragraph. This suggestions seems like overkill with too many potential problems/abuse.

thecrown's strategy was an anomaly for the CC point system, so revamping so much to address so little (which can already be addressed through means already in place) is counter-productive.

[Metsfanmax]

No, this is really a problem with this suggestion. You'll have to get another suggestion through to change the scoreboard. Because under the current scoreboard, this one won't work.

What you call not working, I call a difference in philosophy. I think that if you're good enough to be Conqueror, you should not be able to just go down to major. At least, not easily. I'm surprised that people disagree with that. I can find numerous examples of people posting in Suggestions and C&A to this effect if I need to.

I do concede that this would work much better in a different points system. I'm just not sure if that's a fruitful approach to take because that's a very difficult problem to tackle.

That's apples and oranges. People have to start somewhere.

Yes, but my point is that you need to do more work than just saying "points are created" to prove that the suggestion is a bad idea. As an example, consider that the median score on the scoreboard is around 1250. If we take that as the median score across all players on the site, it means that every time a person joins the site at 1000 and stops playing at 1250 (on average), they are robbing the system of 250 points. This has a long-term deflationary effect. So maybe a little inflation is good for us

I'm just not really sure how this helps anything. I mean, what's the point? The only noted uses for this is to prevent the very small problem of massive point dumping and to protect people who abandon the site. Neither of these is a big benefit and this seemingly creates more problems than it solves, because of the potential abuse. I'm not sure why we're trying to fix one abuse that rarely happens to create an abuse that would happen more often. The abuse that is to be solved can already be dealt with.

The fact that a person can come up with a plethora of abuse scenarios does not mean that any of them are valid, or if they are, that any of them will occur. You're asserting that this abuse would happen more often without justifying it at all.

thecrown's strategy was an anomaly for the CC point system, so revamping so much to address so little (which can already be addressed through means already in place) is counter-productive.

This has nothing to do with TheCrown (and I probably shouldn't have gotten into that debate with agentcom). The fact that that is an anomaly is why I am not worried about the effects of that sort of thing. I'm more worried about the point dumping that happens more much commonly and silently. We don't usually call that point dumping; we call it bad luck. Let me provide an example to show you what this suggestion is really about. My 'actual' score is probably around 2000 (that is, the score where I should equilibrate if I'm playing seriously). A few months ago, I had a nice string of victories that brought me up past 2400. Soon after that, I had a serious streak of bad luck (I play a lot of 1v1s, so this is possible), and my score dropped down below 1500 within the span of like a month. My skill level is not actually 1447, it's much higher than that; but people were playing me at that level, and either not gaining enough points when they won or losing too many when they lost. This happens all the time because of the luck factor in this game, and it's why I'm enthusiastic about this idea. It's just hard to articulate that as a 'benefit' of the suggestion because it happens so frequently and without fanfare.

BTW, I've changed the OP to reflect the 'games played' requirement instead of the 'length of time' requirement. Number of games played at the floor to lower it, is also up for debate of course.

[spiesr]

No, this is really a problem with this suggestion. You'll have to get another suggestion through to change the scoreboard. Because under the current scoreboard, this one won't work.

What you call not working, I call a difference in philosophy. I think that if you're good enough to be Conqueror, you should not be able to just go down to major. At least, not easily. I'm surprised that people disagree with that. I can find numerous examples of people posting in Suggestions and C&A to this effect if I need to.

I do concede that this would work much better in a different points system. I'm just not sure if that's a fruitful approach to take because that's a very difficult problem to tackle.

As long as the points system and overall game here works in such a way that it is realistic for a player to have "Conquerer level" skill in one game type & Major level in another, and that both of those game types are scored together, then that player should be able to go from being Conquer level score to being a Major and there shouldn't be any artificial barriers to that.

[Metsfanmax]

No, this is really a problem with this suggestion. You'll have to get another suggestion through to change the scoreboard. Because under the current scoreboard, this one won't work.

What you call not working, I call a difference in philosophy. I think that if you're good enough to be Conqueror, you should not be able to just go down to major. At least, not easily. I'm surprised that people disagree with that. I can find numerous examples of people posting in Suggestions and C&A to this effect if I need to.

I do concede that this would work much better in a different points system. I'm just not sure if that's a fruitful approach to take because that's a very difficult problem to tackle.

As long as the points system and overall game here works in such a way that it is realistic for a player to have "Conquerer level" skill in one game type & Major level in another, and that both of those game types are scored together, then that player should be able to go from being Conquer level score to being a Major and there shouldn't be any artificial barriers to that.

Since we do have a scoring system where everything is scored together, it's fundamentally meaningless to talk about "Conqueror level skill" on one map and "Major level skill" on another. You have one skill level when it comes to this scoreboard, and one skill level only. No one is Conqueror of Hive. They are Conqueror of CC.

[OliverFA]

Not sure if the Player Score Floor would solve the problem, but the problem does exist, and it's called "Player score not always properly represent player skill".

And this proposal gets close to what could solve this problem, which is having smoother, less drastic, points variation. Perhaps if score did not change so suddenly and evolved more slowly through time the problem would be solved.

The problem is that when a cook wins a single game against a general, that cook gets 100 points, just for a single game. And those 100 points represent 10%, 15% or more of the total cook points. Likewise, the general loses 100 points for a single game, and those points represent 2.5% or more of the total general points. For a single game!

[spiesr]

Since we do have a scoring system where everything is scored together, it's fundamentally meaningless to talk about "Conqueror level skill" on one map and "Major level skill" on another. You have one skill level when it comes to this scoreboard, and one skill level only. No one is Conqueror of Hive. They are Conqueror of CC.

For the time that they play that game type primarily (exclusively?). Should they switch to playing something else then their score will adjust to whatever level is sustainable with their ability at the new game type. Trying to restrict that natural transition wouldn't make any sense to me.

[Metsfanmax]

Since we do have a scoring system where everything is scored together, it's fundamentally meaningless to talk about "Conqueror level skill" on one map and "Major level skill" on another. You have one skill level when it comes to this scoreboard, and one skill level only. No one is Conqueror of Hive. They are Conqueror of CC.

For the time that they play that game type primarily (exclusively?). Should they switch to playing something else then their score will adjust to whatever level is sustainable with their ability at the new game type. Trying to restrict that natural transition wouldn't make any sense to me.

You're ignoring the other half of the equation -- was it a "natural" transition for them to get all the way up to Conqueror by playing a single game type exclusively? Of course not. It was an artificial attempt to manipulate the scoreboard -- it's just one that we've all seemingly accepted as a legitimate method to reach Conqueror.

Not sure if the Player Score Floor would solve the problem, but the problem does exist, and it's called "Player score not always properly represent player skill".

And this proposal gets close to what could solve this problem, which is having smoother, less drastic, points variation. Perhaps if score did not change so suddenly and evolved more slowly through time the problem would be solved.

The problem is that when a cook wins a single game against a general, that cook gets 100 points, just for a single game. And those 100 points represent 10%, 15% or more of the total cook points. Likewise, the general loses 100 points for a single game, and those points represent 2.5% or more of the total general points. For a single game!

Less drastic points variation is an absolute key if we want a player's score to reflect their skill. If you can gain several hundred points at once in a BR it really negates the ability of your score to reflect your skill. It's just a tough problem to solve, given the incredible variety of game types and players we have on the site. It's one I want to attack, but also one that will annoy lots of people regardless of which one we end up selecting.

[OliverFA]

Less drastic points variation is an absolute key if we want a player's score to reflect their skill. If you can gain several hundred points at once in a BR it really negates the ability of your score to reflect your skill. It's just a tough problem to solve, given the incredible variety of game types and players we have on the site. It's one I want to attack, but also one that will annoy lots of people regardless of which one we end up selecting.

Because we acknowledge that the problem exist and we also acknowledge that we want a solution that minimizes it instead of getting worse, I think that suggestions like yours should be debated with the intention to get something from the discussion.

As I said, I am not sure that score floor is the solution, but at least you have proposed something, which is far better than saying "No, it won't work", and I think that has to be recognised.

The thing is that everything in the score system favours huge score changes instead of soft, organic, flowing changes. Changes should be in slower increments and more often. Some solutions I can think about:

- Multiply everybody's score by 10. So a cook does not jump several ranks when winning a game, and a general does not lose so many points because of a bad luck strike.
- Reduce points variation by using an exponential scoring system instead of a linear one viewtopic.php?f=4&t=193115
- Introduce game settings less extreme than the current "winner takes it all" and much more fair than terminator.
- And of course, this proposal.