|
|
Post by den1z on Jan 14, 2017 4:22:54 GMT -5
Hi folks, just a quick question please. Am I right to understand that i will be less able to switch mechs around in my hanger now without incurring play penalties? I'm currently quite happily heading to what I hope is a more stable (less clubbers) high silver level. From what I read this is the aim of the changes so Yay. But I do enjoy switching mechs and playing in different set ups sometimes. Will this still be possible or should I just sit back and wait and see?
|
|
|
|
Post by powerengine333 on Jan 14, 2017 10:27:58 GMT -5
Hello !!
Tell me how to send my robots that my arms TO ANOTHER player ;;;;
Thanks
|
|
|
|
Post by Muhlakai on Jan 14, 2017 11:18:05 GMT -5
Folks are talking about the Elo scores like it penalizes lower-tier bots and imply that that's a bad thing. The reality is that an Elo-based system will put you exactly where you are competitive with the hangar you choose to play. If you want to play lower-tier bots then the Elo system will adjust your score and move bigger, heavier bots out of your tier. If you start throwing bigger bots in your hangar it's going to float you upwards until you're seeing people who play at the same level as you with whatever hangar they're fielding. These are GOOD THINGS.
Want the highest score? Then yes, you'll need 12/12s. ...but fundamentally, that's as silly as clubbing just to get a 100% win rate. Choose what you want to play and the Elo WILL take you to a place where you get just the right challenge with that hangar. The Elo should let you play Triple Zenit Furies and Arty Nats all day long and WIN with those if that's what you want. If you run 3 Galahads, a Lance and a Rhino then the Elo will float you around to where you have lots of challenge with those bots, instead. Are you really good? Well, then you'll see opponents with heavier bots than you, but they will have about the same chance of winning as you, rather than blowing you out just because they're heavier.
My only concern is that it could be much harder to tier with players who are at a different level. For starters, Elo scores don't typically shift too fast and as a rule you don't want players to be able to arbitrarily lower their own score on purpose (because clubbing). It might be much harder to swap between silver and trash gold, for instance, because after playing trash gold you'll have to be losing games until you land back in he right spot with your Silver hangar. (...and if your trip to Gold improved your skills you might not have weapons/bots low enough to take you as far down as you used to be.) It was mentioned that hangar score would still play a role, but hat might be as simple as a bump to the calculated Elo score, which would help, but probably wouldn't be perfect.
Overall, though, it shouldn't be the clusterfiretruck that everyone seems to fear. Want to face other silver-tier bots? Then you don't WANT a top Elo score. The thing you're complaining about will give you EXACTLY what you're asking for.
|
|
|
|
Post by Golden Sabre on Jan 14, 2017 11:26:22 GMT -5
I'm curious how the squadding is going to work. Say all of my squad is what we would call now, bronze and I squad with them as a gold......how does this new hybrid MM sort that that math and make it a 50/50% for everyone ?
|
|
|
|
Post by ShutUpAndSmokeMyWeed on Jan 14, 2017 11:43:41 GMT -5
I'm curious how the squadding is going to work. Say all of my squad is what we would call now, bronze and I squad with them as a gold......how does this new hybrid MM sort that that math and make it a 50/50% for everyone ? In that case it's sort of impossible to make the match fair for everyone, so it'll just take the highest player's rating and use that for matchmaking. |
|
|
|
Post by Golden Sabre on Jan 14, 2017 11:45:22 GMT -5
Yikes
|
|
Deleted
Deleted Member
Posts: 0
Karma:
|
Post by Deleted on Jan 14, 2017 15:05:07 GMT -5
Folks are talking about the Elo scores like it penalizes lower-tier bots and imply that that's a bad thing. The reality is that an Elo-based system will put you exactly where you are competitive with the hangar you choose to play. If you want to play lower-tier bots then the Elo system will adjust your score and move bigger, heavier bots out of your tier. If you start throwing bigger bots in your hangar it's going to float you upwards until you're seeing people who play at the same level as you with whatever hangar they're fielding. These are GOOD THINGS. Want the highest score? Then yes, you'll need 12/12s. ...but fundamentally, that's as silly as clubbing just to get a 100% win rate. Choose what you want to play and the Elo WILL take you to a place where you get just the right challenge with that hangar. The Elo should let you play Triple Zenit Furies and Arty Nats all day long and WIN with those if that's what you want. If you run 3 Galahads, a Lance and a Rhino then the Elo will float you around to where you have lots of challenge with those bots, instead. Are you really good? Well, then you'll see opponents with heavier bots than you, but they will have about the same chance of winning as you, rather than blowing you out just because they're heavier. My only concern is that it could be much harder to tier with players who are at a different level. For starters, Elo scores don't typically shift too fast and as a rule you don't want players to be able to arbitrarily lower their own score on purpose (because clubbing). It might be much harder to swap between silver and trash gold, for instance, because after playing trash gold you'll have to be losing games until you land back in he right spot with your Silver hangar. (...and if your trip to Gold improved your skills you might not have weapons/bots low enough to take you as far down as you used to be.) It was mentioned that hangar score would still play a role, but hat might be as simple as a bump to the calculated Elo score, which would help, but probably wouldn't be perfect. Overall, though, it shouldn't be the clusterfiretruck that everyone seems to fear. Want to face other silver-tier bots? Then you don't WANT a top Elo score. The thing you're complaining about will give you EXACTLY what you're asking for. The lack of differentiation between skill and gear is that part that bothers me. It's not entirely enjoyable being in games where most opponents have heavier bots and ?poo-poo?tier win percentages than you. If you have any measure of skill in this game, it will be a lot easier for the MM to find opponents who out mech you rather than finding opponents with similar gear and skill. If I have to work every shot and angle to win, while my opponent can just blast away blindly with superior firepower and durability, yet we both stand to gain the same reward, that just doesn't sit well with me. The time it takes for the system to reevaluate changes in your hangar in a pure Elo scheme is also a sore point. As you pointed out, sudden downward changes in ability weren't really envisioned when the methodology was developed, so it is going to struggle if you decide you are going to run all lights instead of 9/9 heavies. Similarly, it can be gamed upwardly by playing with a low level hangar while doing your upgrades in the background, then bringing in those much improved weapons/bots in blocks to make big jumps in overall power. Really the underlying system expects either a constant level of ability or a steady progression in skill until in reaches its potential peak. This is fine for the part of the game this is skill based, but material assets that can be switched around at will are entirely another story. |
|
|
|
Post by ShutUpAndSmokeMyWeed on Jan 14, 2017 15:42:09 GMT -5
Folks are talking about the Elo scores like it penalizes lower-tier bots and imply that that's a bad thing. The reality is that an Elo-based system will put you exactly where you are competitive with the hangar you choose to play. If you want to play lower-tier bots then the Elo system will adjust your score and move bigger, heavier bots out of your tier. If you start throwing bigger bots in your hangar it's going to float you upwards until you're seeing people who play at the same level as you with whatever hangar they're fielding. These are GOOD THINGS. Want the highest score? Then yes, you'll need 12/12s. ...but fundamentally, that's as silly as clubbing just to get a 100% win rate. Choose what you want to play and the Elo WILL take you to a place where you get just the right challenge with that hangar. The Elo should let you play Triple Zenit Furies and Arty Nats all day long and WIN with those if that's what you want. If you run 3 Galahads, a Lance and a Rhino then the Elo will float you around to where you have lots of challenge with those bots, instead. Are you really good? Well, then you'll see opponents with heavier bots than you, but they will have about the same chance of winning as you, rather than blowing you out just because they're heavier. My only concern is that it could be much harder to tier with players who are at a different level. For starters, Elo scores don't typically shift too fast and as a rule you don't want players to be able to arbitrarily lower their own score on purpose (because clubbing). It might be much harder to swap between silver and trash gold, for instance, because after playing trash gold you'll have to be losing games until you land back in he right spot with your Silver hangar. (...and if your trip to Gold improved your skills you might not have weapons/bots low enough to take you as far down as you used to be.) It was mentioned that hangar score would still play a role, but hat might be as simple as a bump to the calculated Elo score, which would help, but probably wouldn't be perfect. Overall, though, it shouldn't be the clusterfiretruck that everyone seems to fear. Want to face other silver-tier bots? Then you don't WANT a top Elo score. The thing you're complaining about will give you EXACTLY what you're asking for. The lack of differentiation between skill and gear is that part that bothers me. It's not entirely enjoyable being in games where most opponents have heavier bots and ?poo-poo?tier win percentages than you. If you have any measure of skill in this game, it will be a lot easier for the MM to find opponents who out mech you rather than finding opponents with similar gear and skill. If I have to work every shot and angle to win, while my opponent can just blast away blindly with superior firepower and durability, yet we both stand to gain the same reward, that just doesn't sit well with me. The time it takes for the system to reevaluate changes in your hangar in a pure Elo scheme is also a sore point. As you pointed out, sudden downward changes in ability weren't really envisioned when the methodology was developed, so it is going to struggle if you decide you are going to run all lights instead of 9/9 heavies. Similarly, it can be gamed upwardly by playing with a low level hangar while doing your upgrades in the background, then bringing in those much improved weapons/bots in blocks to make big jumps in overall power. Really the underlying system expects either a constant level of ability or a steady progression in skill until in reaches its potential peak. This is fine for the part of the game this is skill based, but material assets that can be switched around at will are entirely another story. The point you make is valid, but your no-skill opponent would also have played longer and logged more matches than you, so it makes sense to have this progression over time as well. Keep in mind that the eventual league system that all of this is leading up to will also have better rewards for higher leagues, so the better you play the faster you'll be able to upgrade and move up through the tiers  Of course the system isn't even live yet, so we still have no clue by how much ratings will update after each match. There should be a "confidence" variable that determines how much your rating gets adjusted after every match (which is the variance of your performance for the statistically-inclined) which is tracked separately for each player and can also change. For example, it could increase after changing your hangar so that your rating adjusts a lot in the next few matches. Yet another solution could be to make a separate casual or unranked mode that doesn't count for anything and uses your raw hangar score for matchmaking. Anyway, that's my take based on looking at the bright side  |
|
|
|
Post by ł⸰§ĦȺĐ◎ŴƧŦḀɌ on Jan 14, 2017 15:56:41 GMT -5
A lot of this reminds me of the incredibly vitriol debates surrounding Overwatch's competitive (and 'pure') Elo ranking system. Due to player pressure, they've had to change it three times so far. In the latest changes they revamped the gameplay modes as well, including an 'arcade' mode with fewer restrictions and more types of matches (1v1, 3v3, etc) and changing their 'quick play' mode to have a lot more sensible restrictions, making it similar to their competitive mode. EVERYTHING has flaws. Humans are not perfect beings, ergo we cannot create perfect things. What saddens me is when people get fixated (the kids call it 'triggered' now-a-days) on the potential flaws in proposed changes to a system. I don't think anybody has enough information to call it as going to Heaven or Hell just yet. Full Disclosure: I main D.Va on XBox 1  |
|
Deleted
Deleted Member
Posts: 0
Karma:
|
Post by Deleted on Jan 14, 2017 16:27:07 GMT -5
I think having defined leagues / brackets / tiers is the simplest solution to pair with an underlying Elo-esque system, so if that's the plan then I'm all good. About a half dozen different ones would probably do the trick, covering the progression from neophyte to maxed hangar... with rewards increasing accordingly. That way I've got the option of trying to be one of the top dogs in my current "weight class" or move up to the next in search of a higher max payout (with the corresponding increase in challenge).
|
|
|
|
Post by den1z on Jan 14, 2017 20:09:38 GMT -5
Folks are talking about the Elo scores like it penalizes lower-tier bots and imply that that's a bad thing. The reality is that an Elo-based system will put you exactly where you are competitive with the hangar you choose to play. If you want to play lower-tier bots then the Elo system will adjust your score and move bigger, heavier bots out of your tier. If you start throwing bigger bots in your hangar it's going to float you upwards until you're seeing people who play at the same level as you with whatever hangar they're fielding. These are GOOD THINGS. Want the highest score? Then yes, you'll need 12/12s. ...but fundamentally, that's as silly as clubbing just to get a 100% win rate. Choose what you want to play and the Elo WILL take you to a place where you get just the right challenge with that hangar. The Elo should let you play Triple Zenit Furies and Arty Nats all day long and WIN with those if that's what you want. If you run 3 Galahads, a Lance and a Rhino then the Elo will float you around to where you have lots of challenge with those bots, instead. Are you really good? Well, then you'll see opponents with heavier bots than you, but they will have about the same chance of winning as you, rather than blowing you out just because they're heavier. My only concern is that it could be much harder to tier with players who are at a different level. For starters, Elo scores don't typically shift too fast and as a rule you don't want players to be able to arbitrarily lower their own score on purpose (because clubbing). It might be much harder to swap between silver and trash gold, for instance, because after playing trash gold you'll have to be losing games until you land back in he right spot with your Silver hangar. (...and if your trip to Gold improved your skills you might not have weapons/bots low enough to take you as far down as you used to be.) It was mentioned that hangar score would still play a role, but hat might be as simple as a bump to the calculated Elo score, which would help, but probably wouldn't be perfect. Overall, though, it shouldn't be the clusterfiretruck that everyone seems to fear. Want to face other silver-tier bots? Then you don't WANT a top Elo score. The thing you're complaining about will give you EXACTLY what you're asking for. Thanks Muhlakai, at the end of the day I want challenging matches and look forward to this element of change . I really don't get the shooting fish in a barrel mentality but each to their own. But I do like changing mechs and was concerned about that. |
|
|
|
Post by Muhlakai on Jan 14, 2017 20:34:37 GMT -5
The lack of differentiation between skill and gear is that part that bothers me. Thanks Muhlakai, at the end of the day I want challenging matches and look forward to this element of change . I really don't get the shooting fish in a barrel mentality but each to their own. But I do like changing mechs and was concerned about that. There are two methods I suspect they'll choose between to calculate changes in hangars: OPTION A: They compute a hangar score like they've done but they divide that score by 6 or 7 and add it to the Elo. A difference of 100 in Elo means that the higher score should win 50% more often. So they'll set the numbers so that a bot level or two and a few weapon levels is 100 or so and that'll add or subtract from your total Elo as you swap your hangar around. OPTION B: They probably won't start this way, but the really slick way to account for hangar changes is to give each weapon and each frame (at each level) an Elo score (game-wide) based on its prevalence on winning and losing teams. Each weapon copy and frame copy could add its Elo to a hangar score, which (like above) is divided by a number (say, the total number of possible weapon slots and bots in that hangar) and then divided by an additional coefficient (to make it a modifier, and not the tail that wags the dog). These numbers could be updated automatically each week and the score changes could automatically be populated in a score change report that we're shown after login. The reason they're unlikely to adopt OPTION B initially is that until a significant pile of data is compiled those scores are likely to juggle around dramatically at first. Long-term, however, the scores would organically adjust after every balance change and reduce the work required by the devs. It would also highlight weapons or bots that were OP but automatically adjust the MM to account for the imbalance, which would often correct the imbalance all by itself. |
|
|
|
Post by Golden Sabre on Jan 14, 2017 20:49:42 GMT -5
So, in a nutshell....with the new system there are 3 options.
1: You go beast mode
2: You get taken behind the woodshed
3: (most oftenly) It's a close battle to the sweet or bitter end.
Does this sound right ?
|
|
|
|
Post by silverstone21 on Jan 14, 2017 21:38:00 GMT -5
Fujin got a big upgrade. Carnage has the same price and much lower hp and shield. And having the same type of shield and increase use of plasma it should be upgraded
|
|
|
|
Post by silverstone21 on Jan 14, 2017 22:02:56 GMT -5
Or at least put his shield in a way that someone is shooting some meters far away from him and his hitting the shield, but there is no carnage behind it. Just air lol
|
|
|
|
Post by Golden Sabre on Jan 14, 2017 22:07:46 GMT -5
Carnage can also run, fujin can barely beat the piece of grass growing next to it.
|
|
|
|
Post by Muhlakai on Jan 14, 2017 23:23:10 GMT -5
So, in a nutshell....with the new system there are 3 options. 1: You go beast mode 2: You get taken behind the woodshed 3: (most oftenly) It's a close battle to the sweet or bitter end. Does this sound right ? I suspect that most matches won't necessarily be all that close, although more of them will be close than in the current system. It'll be more of a seesaw between wins and losses. You also shouldn't be taken behind the woodshed very often because clubbers would have to lose a bunch of matches to win a bunch of matches, which makes it silly to do rather than sitting in a reasonable tier. ...but we'll see. |
|
|
|
Post by ivx on Jan 15, 2017 0:12:25 GMT -5
Yet another solution could be to make a separate casual or unranked mode that doesn't count for anything and uses your raw hangar score for matchmaking. Anyway, that's my take based on looking at the bright side This too. They could make a simple Casual game mode like Hearthstone/Elder Scrolls: Legends or other card games, and keep the League system at the same time. |
|
|
|
Post by den1z on Jan 15, 2017 4:32:21 GMT -5
So, in a nutshell....with the new system there are 3 options. 1: You go beast mode 2: You get taken behind the woodshed 3: (most oftenly) It's a close battle to the sweet or bitter end. Does this sound right ? I suspect that most matches won't necessarily be all that close, although more of them will be close than in the current system. It'll be more of a seesaw between wins and losses. You also shouldn't be taken behind the woodshed very often because clubbers would have to lose a bunch of matches to win a bunch of matches, which makes it silly to do rather than sitting in a reasonable tier. ...but we'll see. It's all quite exciting. Can't wait |
|
|
|
Post by Loop_Stratos on Jan 15, 2017 10:27:57 GMT -5
I am personally very hype. The Gareth speed buff(there are plenty of times i felt fast yet slow), The new MM(I am an average-above average pilot, so it should help), and the thought of being a sniper/archer and winning due to how the MM will work.
|
|
|
|
Post by debastec on Jan 15, 2017 13:13:08 GMT -5
Estoy curioso cómo el escuadras se va a trabajar. Dicen todos mi equipo es lo que llamaríamos ahora, el bronce y el equipo de I con ellos como un oro ...... ¿cómo esta nueva especie híbrida que MM que las matemáticas y que sea del 50/50% para todo el mundo? |
|
|
|
Post by debastec on Jan 15, 2017 13:14:12 GMT -5
HELLO
Any next novelty in war robots?
|
|
|
|
Post by Muhlakai on Jan 15, 2017 14:30:29 GMT -5
HELLO Any next novelty in war robots? HELLO Yes. Any. |
|
karljackson
Destrier
Posts: 120
Karma: 28
Platform: Android
League: Champion
Server Region: North America
|
Post by karljackson on Jan 16, 2017 7:10:54 GMT -5
Yeah, if you've done well in low gold, you can hang up ever dropping to silver again. You'll face almost the same matches. Been that way since the Halloween MM updates.
|
|
|
|
Post by frunobulax on Jan 16, 2017 8:21:43 GMT -5
If the hangar score is still any part of the new system, I really hope that they make changes to the calculation. I use 2 robots in 95% of my matches, and 3 robots in maybe 80%. (4th and 5th robot do see less playing time.) So, the "correct" hangar score would be 100% value of the strongest robot plus 95% of the second strongest plus 80% of the third strongest plus 50% of robots #4 and #5, or something like this.
|
|
|
|
Post by Muhlakai on Jan 16, 2017 11:39:22 GMT -5
If the hangar score is still any part of the new system, I really hope that they make changes to the calculation. I use 2 robots in 95% of my matches, and 3 robots in maybe 80%. (4th and 5th robot do see less playing time.) So, the "correct" hangar score would be 100% value of the strongest robot plus 95% of the second strongest plus 80% of the third strongest plus 50% of robots #4 and #5, or something like this. They don't need to do that at all. Let's say you swapped in Ecu Cossacks for the last two bots. Your hangar score would go down but your Elo would go up as a result of you winning more after dropping down. Overall you would end up slightly lower than where originally because those two Cossacks would be reducing your ability to win just slightly. |
|
|
|
Post by petevb on Jan 16, 2017 13:11:15 GMT -5
I'd like the Elo score/ rating to be both visible and the goal.
The concern is players will game the system. Example: win Player of the Hour by intentionally crashing your ELO beforehand, then getting serious during the event itself. The system must be durable against manipulation.
This seems possible if the rating itself is the goal. Bots can have their own ELO ratings so statistically strong hangers (i.e. mag Gepards) weigh fairly on a player's composite score. This should end seal-clubbing and make for much closer matches, but it will by its nature bring all win percentages towards 50%. Thus it seems inherently unfair to strong players unless you make the Elo rating itself both the goal and the reward.
I'd like to see gold rewards for higher ratings, and I'd insure that a well played match increases a player's rating even in a loss. There must be a reason to fight a losing battle, otherwise good players will suffer and quitters will ruin the game even more than they already do.
$.02
|
|
|
|
Post by Muhlakai on Jan 16, 2017 14:00:59 GMT -5
I'd like the Elo score/ rating to be both visible and the goal. The concern is players will game the system. Example: win Player of the Hour by intentionally crashing your ELO beforehand, then getting serious during the event itself. The system must be durable against manipulation. This seems possible if the rating itself is the goal. Bots can have their own ELO ratings so statistically strong hangers (i.e. mag Gepards) weigh fairly on a player's composite score. This should end seal-clubbing and make for much closer matches, but it will by its nature bring all win percentages towards 50%. Thus it seems inherently unfair to strong players unless you make the Elo rating itself both the goal and the reward. I'd like to see gold rewards for higher ratings, and I'd insure that a well played match increases a player's rating even in a loss. There must be a reason to fight a losing battle, otherwise good players will suffer and quitters will ruin the game even more than they already do. $.02 +1 to a visible score. It hadn't even occurred to me that they would ever hide it, but based on the current system, who knows. Hopefully they would show not only the total, but also how much of that ScoreIt based on your hangar vs your skill. In any Elo system I can't see win rate being very important anymore. The point of the system is precisely to give you matches with people whose (skill+bots) is closest to your (skill+bots). I'm not certain how that ends up being unfair to people like you and I with high win rates. Rather, our scores will simply be higher and we'd talk about our (unmodified) skill portion of the score rather than our win rate. In most other games with multiple players on each side the Elo usually looks at each individual player's score compared to the average, modified by a comparison of that player's Elo and the average Elo of the team. So you can win but not gain much or lose and gain a bunch, but really only under specific conditions, like you'd expect. It seems reasonable that they *could* award Au for a performance that was significantly better than the average, but I'm not sure that wouldn't encourage clubbing. You'll probably see some folks crashing their score for things like PotH, (which I think would have to be done away with anyhow) but now that Au can be gifted anyhow they could accomplish the same thing easier by starting a new account. A weekly reward system for having kept your score above certain benchmarks all week will also discourage clubbing on a main account, but again, Au gifting still enables new-account clubbing. |
|
Deleted
Deleted Member
Posts: 0
Karma:
|
Post by Deleted on Jan 16, 2017 15:03:41 GMT -5
I'd like the Elo score/ rating to be both visible and the goal. The concern is players will game the system. Example: win Player of the Hour by intentionally crashing your ELO beforehand, then getting serious during the event itself. The system must be durable against manipulation. This seems possible if the rating itself is the goal. Bots can have their own ELO ratings so statistically strong hangers (i.e. mag Gepards) weigh fairly on a player's composite score. This should end seal-clubbing and make for much closer matches, but it will by its nature bring all win percentages towards 50%. Thus it seems inherently unfair to strong players unless you make the Elo rating itself both the goal and the reward. I'd like to see gold rewards for higher ratings, and I'd insure that a well played match increases a player's rating even in a loss. There must be a reason to fight a losing battle, otherwise good players will suffer and quitters will ruin the game even more than they already do. $.02 +1 to a visible score. It hadn't even occurred to me that they would ever hide it, but based on the current system, who knows. Hopefully they would show not only the total, but also how much of that ScoreIt based on your hangar vs your skill. In any Elo system I can't see win rate being very important anymore. The point of the system is precisely to give you matches with people whose (skill+bots) is closest to your (skill+bots). I'm not certain how that ends up being unfair to people like you and I with high win rates. Rather, our scores will simply be higher and we'd talk about our (unmodified) skill portion of the score rather than our win rate. In most other games with multiple players on each side the Elo usually looks at each individual player's score compared to the average, modified by a comparison of that player's Elo and the average Elo of the team. So you can win but not gain much or lose and gain a bunch, but really only under specific conditions, like you'd expect. It seems reasonable that they *could* award Au for a performance that was significantly better than the average, but I'm not sure that wouldn't encourage clubbing. You'll probably see some folks crashing their score for things like PotH, (which I think would have to be done away with anyhow) but now that Au can be gifted anyhow they could accomplish the same thing easier by starting a new account. A weekly reward system for having kept your score above certain benchmarks all week will also discourage clubbing on a main account, but again, Au gifting still enables new-account clubbing. Visible score is a definite positive. Anything that improves transparency rather than having players trying to guess what's going on under the hood or allowing players (particularly new ones) to make decisions that appear good on the surface have have substantial negative implications (e.g. adding that 6/5 Nat to your twin 3/3 Cossack hangar under the current MM). Under an Elo system, rating definitely needs to replace win rate as the central metric. That also means rewards would need to be more strongly tied to rating rather than the simple win/loss approach of the current system that encourages players to look for lopsided matchups in their favor. You definitely need some mechanism to prevent tanking as a viable manipulation of rewards - but I think they could tie this to solving the greater issue of players who regularly drop games as intentional tanking isn't that appealing if you actually have to play each match (rather than just repeating a "join-quit-join-quit" sequence). |
|
|
|
Post by frunobulax on Jan 16, 2017 18:13:16 GMT -5
If the hangar score is still any part of the new system, I really hope that they make changes to the calculation. [...] They don't need to do that at all. Let's say you swapped in Ecu Cossacks for the last two bots. Your hangar score would go down but your Elo would go up as a result of you winning more after dropping down. Overall you would end up slightly lower than where originally because those two Cossacks would be reducing your ability to win just slightly. Well, no, not really Just for the sake of the argument, assume that the hangar score counts 90% and the ELO 10%. Then you would lose a lot of battles against equalized hangars, because the ELO won't compensate enough. My first choice would be to make everything ELO-based, because this way you don't have to worry about correctness of hangar score. But if you do use hangar score at all, you should try to make it as precise as possible. |
|
