|
Post by Dredd77 on May 24, 2017 23:16:53 GMT -5
|
|
|
Post by WE034 on May 24, 2017 23:24:44 GMT -5
When do we get to test the "paddle thwacker" on the TS? Thanks for the post!
|
|
Deleted
Deleted Member
Posts: 0
Karma:
|
Post by Deleted on May 24, 2017 23:59:29 GMT -5
As far as taking exactly the same levels is concerned, yes, the average formula wouldn't work.
However, assigning a value might work. Every 10k of armour, 1pt. Weapons round to the nearest 1k / 10 so 1k = 10, 1949 = 19, 2853 = 29.
Laverne and Shirley Armour - 18pts Laverne and Shirley weapons - 36pts Total - 54pts
Cagney and Lacey Armour - 20pts Cagney and Lacey Weapons - 40pts Total - 60pts
Giving a variable +/-5 for a match up, they wouldn't meet.
Special abilities, depending on the ability, would add on points. Shields would be assigned a value for armour as well.
Probably change the meta and something would be the next go to bot. However, it should create relatively even matches.
|
|
|
Post by moody on May 25, 2017 0:35:32 GMT -5
I think hangar rating more than hangar average is more often floated.
We know average doesn't work. That's how gep clubbers used to manage with 2 lev 1 and a lev 12 magnum.
|
|
Deleted
Deleted Member
Posts: 0
Karma:
|
Post by Deleted on May 25, 2017 0:51:24 GMT -5
I think by simply adding hard caps to each league preventing entry of certain gear levels into each league they'd solve a lot of player issues.
But as for playing lower bots as a higher league pilot, it just can't be done with the league point based mm, and so I think the only future option will be in their non competitive modes.
|
|
|
Post by llama4president on May 25, 2017 1:09:32 GMT -5
Average hangar by using just lvls wouldn't work, unless they would be adapted counting their exponential progression, and putting some additional rules to avoid exploiting AND possible abuses.
as first points i'd take:
highest weapon equipped as baseWpnLvl. highest bot in hangar as baseBotLvl.
other bots/weapons lower than 15~20% of these base values, would be counted as BaseLvl-15~20%
for Weapons i would count lvls as per your exponential Chart:
lvl 1: 1.0 lvl 2: 1.1 lvl 3: 1.21 lvl 4: 1.331 lvl 5: 1.464 lvl 6: 1.611 lvl 7: 1.772 lvl 8: 1.949 lvl 9: 2.144 lvl 10: 2.358 lvl 11: 2.853
multipliers:
light weapon: x1.0 med weapon: x1.5 heavy weapon: x2.0
Ecu Counted as a normal Light weapon. Ancile Counted as a normal Heavy weapon.
Bots:
Exponential lvl progression similar to the Weapon Chart, and individual multiplier applied per Bot Category[OPTIONAL].
Light x 1 Med x 1.25 Heavy x 1.5
End calculation would sum up:
#1Bot: (sum of all weaponLVLs)+(BotLvl x BotCategoryMultiplier) #2Bot: " " " " " " #3Bot: etc etc
Hangar Lvl would be a sum up of each bot, and in case of bots or weapons lower than -15~20% it would sum them as they were just -15~20% less than the base bot.
This way weapons would have a way heavier weight in hangar judgement (3 or 4 more times more important than Bot Lvl). Bot category multiplier is still something that leaves me some doubts , and could be taken off completely from the equations, as the heavier they are, generally they pack more weapons that would still balance out the general hangar level. So a variation would be totally without any BotCategoryMultiplier.
Just wrote this on rush, quite certainly better solutions could be considered.
EDIT1:
Possible Problems or Abuses of this kind of Calculations:
1)Hangar slots abuse: Not Buying Hangar slots would give an unfair advantage. a 12/12 hangar with 2 slots would be equated with a roughly estimation to a 10/10~9/9 hangar with 3 slots, and a 8/8~7/7 with 4 slots, and a 5/5~6/6 with 5 slots. Solution: If hangar has more than 3 slots, count as having 3 hangar slots, and discarding the lower bots. If 2 slots or less, equate as a 3 hangar slot, counting lowest bot in hangar as their 2nd/3rd.
2)People would tend play with lower level hangars: Playing with lower lvl hangars would generate faster battles, and easier access to gold per battle determined by higher skill. Solution: Make Silver earning scale with HangarLvl, and make macro categories for Gold Earning to entice playing with maxed hangars for higher earnings.
3)Incompatible with the Actual League system. It would bypass totally the league system, so it would become redundant. Solution1: scrap the League system as matching sytem, but use it just as Activity per Month for Prizes. Solution2: make a separate mode that is hangar based, and doesn't count for End of Month prizes.
EDIT2: for the sake of balance, some bots should be considered armed with some weapons to justify their corresponding ability:
carnage: ancile (x0.5 ?) gareth: ecu lvl as bot lvl galahad: ecu lvl as bot lvl x1.25 Lancelot: ecu lvl as bot lvl Fujin: ancile lvl as bot lvl x0.5 Raijin: ecu lvl as bot lvl stalker: stealth = a fixed lvl 4 light weapon? rogatka: jump = a fixed lvl 4 light weapon? cossack: jump = a fixed lvl 2 light weapon? griffin: jump = a fixed lvl 4 light weapon? Rhino: ecu lvl as bot lvl
Wild West Bots: modifier for summed weapons x0.75?
|
|
|
Post by moody on May 25, 2017 2:02:59 GMT -5
A balanced hangar method could be used but you need to balance out what each slot is worth.
Weighting the slots might work. First 100 second 90 third 80 fourth 70 fifth 60. That way a 4 slot hangar would be about 15% less value than a 5 slot hangar.
|
|
|
Post by Tatamat on May 25, 2017 2:04:21 GMT -5
Aye, we know the problem. Is it the only one the average hangar calculation has? 'Cause this one is easy to fix, ain't it? It's not much difference for server to calculate exponential average instead of linear one. The only discussion would be about the basis for exponential function - weapons go by 10 %, bots by 6-6.5%. However, anything in this range would probably work. Another problem to cope with (and a worse one) is balancing bots in their category. Based on Forum discussion, I'd for example assume that Gareth can't be taken with the same weight as other light bots, same for Galahad & Carnage among mediums. IMO it was THIS and not ne non-linearity that has lead Pix to performance-based rating. I remember a piece of argument when they advocated the new MM. It was like "And we realized that we need to put MM penalty on Gepard and Gareth and then Orkan etc. and after some time the MM was just one heap of patches one on another" EDIT: As it took me quite long to write this port, llama has posted something very similar in the meantime.
|
|
|
Post by carnage on May 25, 2017 2:16:31 GMT -5
I don't think it's so much of a big deal if a 2/12 weapons has a reasonable advantage over a 7/7 weapon. The 2/12 setup costs more, and takes more time to upgrade so a little advantage should not be seen as a system weakness. Also it is much easier for the 7/7 to change weapons, since the investment has been much smaller than on a 2/12. So personally, I'd totally disagree to say this disqualify this MM as a possibility.
All that said, I said many time this is not the best option. Power can be very different from two different weapons, just like a level 7 Lancelot is a better robot than a level 7 Leo. This is actually the real problem and this is why the best option is to base hangar on VALUES, and then to create leagues based on those values. It works wonderfully in many games, for some reasons War Robots players like to imagine crazy things instead but a value-based MM is and will always be the best option for this kind of game.
|
|
|
Post by Tatamat on May 25, 2017 3:19:56 GMT -5
I don't think it's so much of a big deal if a 2/12 weapons has a reasonable advantage over a 7/7 weapon. The 2/12 setup costs more, and takes more time to upgrade so a little advantage should not be seen as a system weakness. Also it is much easier for the 7/7 to change weapons, since the investment has been much smaller than on a 2/12. So personally, I'd totally disagree to say this disqualify this MM as a possibility. The amoung of time and money you've invested in your hangar should not affect MM, i.e. the argument is void. Claiming that stronger 2/12 shall be paired with weaker 7/7 because 2/12 has invested more is similar to claim that someone who invested 1200 Au in Gepard should be paired with Destriers and Cossacks. We've had that before
|
|
|
Post by carnage on May 25, 2017 3:46:31 GMT -5
Tatamat, a 2/12 combo has roughly 10% more power than a 7/7. It is not a game-breaking argument at all, especially considering the disadvantages is also have in terms of flexibility (like I said, much easier to change weapons when you are level 7, not the same with level 12). Not sure it would make sense for a player to really go the 2/12 route, since the investment is bigger, the advantage is marginal, and the loss of flexibility is big. Most players would rightfully go the 7/7 route and be very close in terms of power to a 2/12 setup. That's what I would do personally.
What I am saying is that the problem is not here. The problem is, and your example exactly happens to prove my point, that different robot have different power. A level 1 Guepard is stronger than a level 1 Destrier. A level 1 Lancelot is stronger than a level 1 Leo. It's exactly the same with weapons, it just doesn't make sense at all to compare level of items, when those items have by definition very different power.
This is why the only viable system is a value based-MM.
|
|
|
Post by The VVatcher on May 25, 2017 5:36:27 GMT -5
As far as taking exactly the same levels is concerned, yes, the average formula wouldn't work. However, assigning a value might work. Every 10k of armour, 1pt. Weapons round to the nearest 1k / 10 so 1k = 10, 1949 = 19, 2853 = 29. Laverne and Shirley Armour - 18pts Laverne and Shirley weapons - 36pts Total - 54pts Cagney and Lacey Armour - 20pts Cagney and Lacey Weapons - 40pts Total - 60pts Giving a variable +/-5 for a match up, they wouldn't meet. Special abilities, depending on the ability, would add on points. Shields would be assigned a value for armour as well. Probably change the meta and something would be the next go to bot. However, it should create relatively even matches. There's a problem with this. Raijin would be rated much higher than Griffin because it's hp is so much higher. Not sure if the firepower between Raijin and Griffin will balance this out though. So then a better example would be Rhino vs Griffin because they have equal firepower. According to your system Rhino is rated higher when in fact Griffin is the better bot.
|
|
Deleted
Deleted Member
Posts: 0
Karma:
|
Post by Deleted on May 25, 2017 6:24:06 GMT -5
As far as taking exactly the same levels is concerned, yes, the average formula wouldn't work. However, assigning a value might work. Every 10k of armour, 1pt. Weapons round to the nearest 1k / 10 so 1k = 10, 1949 = 19, 2853 = 29. Laverne and Shirley Armour - 18pts Laverne and Shirley weapons - 36pts Total - 54pts Cagney and Lacey Armour - 20pts Cagney and Lacey Weapons - 40pts Total - 60pts Giving a variable +/-5 for a match up, they wouldn't meet. Special abilities, depending on the ability, would add on points. Shields would be assigned a value for armour as well. Probably change the meta and something would be the next go to bot. However, it should create relatively even matches. There's a problem with this. Raijin would be rated much higher than Griffin because it's hp is so much higher. Not sure if the firepower between Raijin and Griffin will balance this out though. So then a better example would be Rhino vs Griffin because they have equal firepower. According to your system Rhino is rated higher when in fact Griffin is the better bot. There's never going to be true balance, it's just a fact. That being said, Rhino vs Griffin perhaps. Personally, I think that if a bot can withstand more punishment than another then it has an advantage. The Griffin is largely more played due to being a lot easier to use, and being able to get out of difficult situations. Rhino takes some working out, rather than the storm in and die that a lot of people use - so likely stop using it for that reason. Obviously, any system that's created would have flaws, but there has to be a better way to match players than what we currently have.
|
|
|
Post by The VVatcher on May 25, 2017 6:34:49 GMT -5
There's a problem with this. Raijin would be rated much higher than Griffin because it's hp is so much higher. Not sure if the firepower between Raijin and Griffin will balance this out though. So then a better example would be Rhino vs Griffin because they have equal firepower. According to your system Rhino is rated higher when in fact Griffin is the better bot. There's never going to be true balance, it's just a fact. That being said, Rhino vs Griffin perhaps. Personally, I think that if a bot can withstand more punishment than another then it has an advantage. The Griffin is largely more played due to being a lot easier to use, and being able to get out of difficult situations. Rhino takes some working out, rather than the storm in and die that a lot of people use - so likely stop using it for that reason. Obviously, any system that's created would have flaws, but there has to be a better way to match players than what we currently have. How about Leo versus Galahad then? With all due respect: 2x Rhino 3x Leo is better than 2x Galahad 3x Lancelot in that system.
|
|
Deleted
Deleted Member
Posts: 0
Karma:
|
Post by Deleted on May 25, 2017 6:40:25 GMT -5
I don't really agree with this because this is assuming that everyone's bot weapons and setups will be the same. Realistically speaking, a guy at 12/12 would have, say a 12/12 Plasma Galahad and more skill. The middle guy, at 7/7 would have a Plasma/DB Griffin at most. Maybe a Thunder Pinata Leo. And if he is paired up with someone with 2/2s (Punisher Cossack), which team do you think will win? It is obviously the team with the 7/7/ and not because the team was unbalanced, but because of the bot type and more people capping beacons. In LQ is the best place to explore this. I have been placed in many 4v4s where I was the only good player on my team and the rest were full of swiss army builds while the enemy had some decently leveled builds. My team meched out and I lost, because there was only guy capping beacons on my team.
And adding on to your story, who is the real winner of the game? Lacey is pissed as hell because she meched out so early. Laverne and Shirley are both pissed because the not only lost to a humiliating 2v1, but also it was an unfair one, where the only deciding factor was, you know, bots and weapons. Cagney is the only one happy because she not only won a 2v1, but she also got to beat up lower leveled players (that is, if Cagney has a clubbing mindset). Had everyone been 7/7, you know, everyone would probably be happy and the two players on the losing team would be slightly upset. It would have been an absolutely
That's the problem with the League system right now; no one is happy and the only guy that is benefited is the guy with a maxed hangar or near maxed hangar.
I'm in a clan with a mix of players anywhere from Silver to Master to a fake Champion status holder (me). I have squadded with many clan members that are ranked lower. Much lower (Champion - Gold 1 squad). We are pulled up to champion in which neither of us are fully prepared for. As a result, it's always a guaranteed loss. And yes, if we were placed in the middle (Expert League statistically, Diamond 1 realistically), we probably would have lost because he would have meched out earlier, leaving one less guy to cap beacons with.
The only way to fix the squadding issue is to go back to the old system. A hangar-based system allows you to evenly match up hangars and you could squad in peace. Yes, you may have a slight edge in skill. But in the end, in War Robots bot/weapons > skill.
|
|
Deleted
Deleted Member
Posts: 0
Karma:
|
Post by Deleted on May 25, 2017 7:00:53 GMT -5
There's never going to be true balance, it's just a fact. That being said, Rhino vs Griffin perhaps. Personally, I think that if a bot can withstand more punishment than another then it has an advantage. The Griffin is largely more played due to being a lot easier to use, and being able to get out of difficult situations. Rhino takes some working out, rather than the storm in and die that a lot of people use - so likely stop using it for that reason. Obviously, any system that's created would have flaws, but there has to be a better way to match players than what we currently have. How about Leo versus Galahad then? With all due respect: 2x Rhino 3x Leo is better than 2x Galahad 3x Lancelot in that system. What do you mean by better? Score lower?
|
|
|
Post by manic502 on May 25, 2017 7:19:12 GMT -5
"Cagney and Lacey" - brooklyn99 fan spotted
|
|
|
Post by ༒ƜƦƛƖƬӇ༒ on May 25, 2017 7:42:09 GMT -5
All this MM talk has me
|
|
|
Post by The VVatcher on May 25, 2017 8:37:13 GMT -5
How about Leo versus Galahad then? With all due respect: 2x Rhino 3x Leo is better than 2x Galahad 3x Lancelot in that system. What do you mean by better? Score lower? It means assigning points per 10K hp is flawed because high hp robots doesn't mean they are the best bots in the game. So the matchmaking will match weaker bots with the meta bots because that system actually says those weaker bots are better because it has more points. Because Galahad has less points than a Leo, a high leveled Galahad will match with lower leveled Leo.
|
|
Deleted
Deleted Member
Posts: 0
Karma:
|
Post by Deleted on May 25, 2017 8:47:17 GMT -5
What do you mean by better? Score lower? It means assigning points per 10K hp is flawed because high hp robots doesn't mean they are the best bots in the game. So the matchmaking will match weaker bots with the meta bots because that system actually says those weaker bots are better because it has more points. Because Galahad has less points than a Leo, a high leveled Galahad will match with lower leveled Leo. I don't see the problem, medium vs heavy. 3 points vs 4points. A maxed medium vs a mid level heavy. I'm just spit balling an idea. It's wide open to improvements.
|
|
|
Post by ᴛʜᴇ ʙɪɢ ᴅʀᴏᴘ on May 25, 2017 10:51:44 GMT -5
Totally appreciate the use of Lenny & Squiggy... H E L L O !
|
|
|
Post by gr3ygh05t on May 25, 2017 11:01:58 GMT -5
Nice post Dredd77. Might I recommend alternate leagues and game modes to fix the uneven squadding issue. Recruit League = Silver bots and weapons only with max Level of 7 Veteran League = WSP bots and weapons allowed with a max level of 9 Ranked League = Normal league with no restrictions Veteran players will now have a use for extra bots and weapons in their hangar and will fix the balance issue when squadding with lower level players.
|
|
|
Post by critter667 on May 25, 2017 11:57:44 GMT -5
When do we get to test the "paddle thwacker" on the TS? Thanks for the post! No kidding! I want a paddle thwacker.
|
|
|
Post by zer00eyz on May 25, 2017 12:13:02 GMT -5
A few things: First great article.... But I think many of you failed to read between the lines: Backing up your random idea with an analysis is what is required for it to be taken seriously, and if you fail to do it someone will come up with the data to rip it to shreds. Even if you do a good job, someone might find the things you missed and shred it. I think that Dredd77 made an interesting choice in what he picked, and I would love to know if he did it on purpose. Hanger average, and the current league system, both have a "feature" that the old MM didn't. The old MM was truly a set of arbitrary rules, that had more arbitrary rules stacked on top of it. Because they were so arbitrary the game got balanced INTO these rules, and the groups and pools they created. Hanger average to some (fairly large) extent, and the leagues to large extent make no such assumptions. There is very little that is arbitrary about them! Because of this, pix has a larger degree of freedom, not only to balance the game, but to let players pick a path. However hanger average is BROKEN in a rather grave and unfixable way, unless you start stacking arbitrary rules on it and those are a non starter, because we end up right back into the old MM.
|
|
|
Post by WE034 on May 25, 2017 13:19:03 GMT -5
A few things: First great article.... But I think many of you failed to read between the lines: Backing up your random idea with an analysis is what is required for it to be taken seriously, and if you fail to do it someone will come up with the data to rip it to shreds. Even if you do a good job, someone might find the things you missed and shred it. I think that Dredd77 made an interesting choice in what he picked, and I would love to know if he did it on purpose. Hanger average, and the current league system, both have a "feature" that the old MM didn't. The old MM was truly a set of arbitrary rules, that had more arbitrary rules stacked on top of it. Because they were so arbitrary the game got balanced INTO these rules, and the groups and pools they created. Hanger average to some (fairly large) extent, and the leagues to large extent make no such assumptions. There is very little that is arbitrary about them! Because of this, pix has a larger degree of freedom, not only to balance the game, but to let players pick a path. However hanger average is BROKEN in a rather grave and unfixable way, unless you start stacking arbitrary rules on it and those are a non starter, because we end up right back into the old MM.
|
|
|
Post by zer00eyz on May 25, 2017 13:48:54 GMT -5
A few things: First great article.... But I think many of you failed to read between the lines: Backing up your random idea with an analysis is what is required for it to be taken seriously, and if you fail to do it someone will come up with the data to rip it to shreds. Even if you do a good job, someone might find the things you missed and shred it. I think that Dredd77 made an interesting choice in what he picked, and I would love to know if he did it on purpose. Hanger average, and the current league system, both have a "feature" that the old MM didn't. The old MM was truly a set of arbitrary rules, that had more arbitrary rules stacked on top of it. Because they were so arbitrary the game got balanced INTO these rules, and the groups and pools they created. Hanger average to some (fairly large) extent, and the leagues to large extent make no such assumptions. There is very little that is arbitrary about them! Because of this, pix has a larger degree of freedom, not only to balance the game, but to let players pick a path. However hanger average is BROKEN in a rather grave and unfixable way, unless you start stacking arbitrary rules on it and those are a non starter, because we end up right back into the old MM. Who knows kung fu?
|
|
|
Post by ᴛʜᴇ ʙɪɢ ᴅʀᴏᴘ on May 27, 2017 11:51:24 GMT -5
Paddle-Wakers are wayyyy op. Im seeing ehtire hangers of those things.
I dont know Kung-Fu, but I do know Cra-zay!
Get on up... Get involved.
|
|
|
Post by Firebeard on May 28, 2017 5:47:41 GMT -5
The Op-Ed assumes a completely random Match-up. The purpose of the "Average Hangar" is to keep Players in their respective Leagues and to avoid incidents such as that described.
In the "Average Hangar" system, Cagney would remain in Champion League, Lacy would be kept in Bronze League while Lavern and Shirley would be in the same League(s) and Matched against each other as Opponents.
Also, the Op-Ed used a different method to solve for "Average" than what has been proposed.
|
|
|
Post by make7upyours on May 30, 2017 16:09:09 GMT -5
Instead of hangar averages, I think each stage should have a level cap. Like level 7 for silver.
In this scenario, if a high level player were playing with a low level friend, high level would be regulated to level cap.
May not be a perfect solution, but it's an improvement.
And also, regarding Dredd's example. What if instead of matching based on level, match on the performance instead(actual damage/health/sped).
|
|