Leaderboard Not Accurate
#1
Posted 02 November 2017 - 07:27 AM
had 11 games when i started today and i made 3 games with my asn-21. Avg. matchscore was 303
1st game was 650dmg+ 2 kills no death. 380ish matchscore
2nd game was 800dmg+ 2 kills 1 death 502 matchscore
last game was 250 dmg 2 kilsl 1 death 200 matchscore
so i should have 14 games and additional 6 kills and 2 death. and my matchscore should rise since the average of those 3 games is above 350. It fell to 299 - i bet only the last game got counted. i also only got 2 kills and 1 death counted on top of my old games from yesterday.
I however got 3 loot bags.
Its has always been tehs ame with personal stats. There are mechs with hundred of games i didnt play and mechs that i played excluseively over a year with almost no counted games.
#2
Posted 02 November 2017 - 07:30 AM
#3
Posted 02 November 2017 - 07:34 AM
14 games, and the match score is correct given the values you provided: [(303*11)+380+502+200]/14=315.35
The leaderboard updates about 5 minutes after the hour. Wait a smidge.
Edited by Bombast, 02 November 2017 - 07:36 AM.
#4
Posted 02 November 2017 - 10:25 AM
Bombast, on 02 November 2017 - 07:34 AM, said:
14 games, and the match score is correct given the values you provided: [(303*11)+380+502+200]/14=315.35
The leaderboard updates about 5 minutes after the hour. Wait a smidge.
Yeah i know all that, i waited over an hour and had lunch to make sure that is not the problem.
What you see there are two additional games i played after lunch. Those recorded corrently It should be 16 by the time you looked it up. I will start playing in half an hour again and i will write down the number of games i played and will check tomorrow wether they have been recorded correctly or not.
On a side note - the match score seems to fit as you calculated. Could it be that the updates are very very slow lately? What if the two missing games are from after the time you looked at it. I wrote the opening post after lunch while dead in a game and then had a very quick game afterwards. Perhaps i am hitting the times between updates here and it not 5 minutes anymore.
Another idea? could the forum and the game be out of sync?
Ahh damnit so manny possibilities.
Edited by Cara Carcass, 02 November 2017 - 10:32 AM.
#5
Posted 02 November 2017 - 10:38 AM
If you solo drop, your win/loss is going to be random because potatoes. Having a bad team will also lower your damage, kills, KMDD's, match score, and attractiveness to the opposite sex.
Even in team drops, your actual stats get to be pretty random because so many of them are dice rolls. Kills are last hit. High damage requires you to get good crit rolls and be the one lucky enough to blow off a side torso with an arm and a bunch of weapons, and so on.
The leaderboards also don't really reflect useful things like a well placed UAV, narc/tagging/R-spotting by lights, actually achieving objectives and so on.
So ya, life is better if you just pretend they don't exist.
#6
Posted 02 November 2017 - 10:50 AM
Xavori, on 02 November 2017 - 10:38 AM, said:
If you solo drop, your win/loss is going to be random because potatoes. Having a bad team will also lower your damage, kills, KMDD's, match score, and attractiveness to the opposite sex.
Even in team drops, your actual stats get to be pretty random because so many of them are dice rolls. Kills are last hit. High damage requires you to get good crit rolls and be the one lucky enough to blow off a side torso with an arm and a bunch of weapons, and so on.
The leaderboards also don't really reflect useful things like a well placed UAV, narc/tagging/R-spotting by lights, actually achieving objectives and so on.
So ya, life is better if you just pretend they don't exist.
Sure some randomness is there of course. But somebody who consistently does better than others will have better stats. Look at my first games - normally i have a win/loss ratio of about 1.4-1.5 i think. Now its 0.33 but over time it will even out.
Sure there are leachers that stay back for match score. I was jsut curios since i know from looking at my mechs and the times they are played, something is wrong. And it seems as if at first look leaderboardf is messed up too. I iwll take time tonight and wriote down my games to compare them with updated stats.
#7
Posted 02 November 2017 - 11:05 AM
Forum Displays 15.
So the update rates for ingame and forum are different.
#8
Posted 02 November 2017 - 11:23 AM
#9
Posted 02 November 2017 - 11:38 AM
Xavori, on 02 November 2017 - 10:38 AM, said:
If you solo drop, your win/loss is going to be random because potatoes. Having a bad team will also lower your damage, kills, KMDD's, match score, and attractiveness to the opposite sex.
Even in team drops, your actual stats get to be pretty random because so many of them are dice rolls. Kills are last hit. High damage requires you to get good crit rolls and be the one lucky enough to blow off a side torso with an arm and a bunch of weapons, and so on.
The leaderboards also don't really reflect useful things like a well placed UAV, narc/tagging/R-spotting by lights, actually achieving objectives and so on.
So ya, life is better if you just pretend they don't exist.
This isn't true though. The only consistent factor in a player's matches is that player. Over a large number of matches a good player playing solo will end up with a higher WLR, KDR, and MS than a bad player. That's how averages work. Sure, if someone only has a handful of matches their numbers can be skewed by luck. Once you start getting to the hundreds of matches it becomes an increasingly accurate representation of that pilot's overall ability.
Stats don't represent skill is usually the rallying cry of player who have poor stats. Things like tag and narc really don't contribute as much as you think they do and even if they do it will reflected in the WLR if it helps the team to win more often on average (it rarely does).
It amazes me how many people don't understand the basic principles of statistics. A few matches doesn't mean much, but over a large sample size stats can absolutely be used to see trends in player ability.
It's fine not to care about stats, but don't act as if stats aren't a valuable metric for approximating player skill. A player that averages a 2.0 KDR over 400 matches is almost certainly going to be better than one who averages a 0.5 KDR over the same number of matches. You can't say that a person with a 2.1 KDR is strictly better than one with a 2.0 KDR, but with larger differences it's almost always accurate for 95% of the players (really high scores tend to be more likely to be skewed by other things).
I averaged a 2.42 WLR over 113 matches in season 16 solo dropping. Those stats show that my presence on the team makes that team more likely to win. It's not luck, I had to carry a lot of close matches to end up with those results. Similarly, someone with a sub 1 WLR is (on average) actively contributing to their team losing matches.
#10
Posted 02 November 2017 - 11:53 AM
Xavori, on 02 November 2017 - 10:38 AM, said:
If you solo drop, your win/loss is going to be random because potatoes. Having a bad team will also lower your damage, kills, KMDD's, match score, and attractiveness to the opposite sex.
Even in team drops, your actual stats get to be pretty random because so many of them are dice rolls. Kills are last hit. High damage requires you to get good crit rolls and be the one lucky enough to blow off a side torso with an arm and a bunch of weapons, and so on.
The leaderboards also don't really reflect useful things like a well placed UAV, narc/tagging/R-spotting by lights, actually achieving objectives and so on.
So ya, life is better if you just pretend they don't exist.
Just read what Xiphias wrote, I don't think I have to add more than that. Yes, on an individual game basis or small sample size, you can get screwed over by bad teammates/random circumstances.
But do you really believe that someone like Proton has over a 4 WLR while playing almost exclusively in solo queue because he consistently gets better teams than you? It's not because of the fact that's he's the best player in the game and has over 450 match score and 7 KDR despite playing random builds?
And yes, you can contribute beyond match score/kills with UAVs and the such, it's just rare that when you're contributing very little damage and not killing anything that you're still helping your team win more often than lose.
Edited by Mighty Wings, 02 November 2017 - 11:57 AM.
#11
Posted 02 November 2017 - 12:03 PM
#12
Posted 02 November 2017 - 12:17 PM
Xiphias, on 02 November 2017 - 11:38 AM, said:
Not true.My current K/D is 0.7 since I've been playing solo in mostly lights since I got back to the game. Before I quit it was getting close to 3 because I played almost exclusively in organized groups and larger mechs. Stats are an accurate representation of some things, but individual skill isn't one of them.
Edited by The Cyberserker, 02 November 2017 - 12:18 PM.
#13
Posted 02 November 2017 - 12:19 PM
Xiphias, on 02 November 2017 - 11:38 AM, said:
Stats don't represent skill is usually the rallying cry of player who have poor stats. Things like tag and narc really don't contribute as much as you think they do and even if they do it will reflected in the WLR if it helps the team to win more often on average (it rarely does).
It amazes me how many people don't understand the basic principles of statistics. A few matches doesn't mean much, but over a large sample size stats can absolutely be used to see trends in player ability.
It's fine not to care about stats, but don't act as if stats aren't a valuable metric for approximating player skill. A player that averages a 2.0 KDR over 400 matches is almost certainly going to be better than one who averages a 0.5 KDR over the same number of matches. You can't say that a person with a 2.1 KDR is strictly better than one with a 2.0 KDR, but with larger differences it's almost always accurate for 95% of the players (really high scores tend to be more likely to be skewed by other things).
I averaged a 2.42 WLR over 113 matches in season 16 solo dropping. Those stats show that my presence on the team makes that team more likely to win. It's not luck, I had to carry a lot of close matches to end up with those results. Similarly, someone with a sub 1 WLR is (on average) actively contributing to their team losing matches.
No.
There is no amount of skill that can make up for a team that doesn't focus fire, scatters, and can't hit the broad side of the barn. Perhaps if we drop to 8v8 one player might be enough to tip the odds, but again, at only 12.5% of the team, it's very iffy.
And statistics blah blah blah. You want math, I can totally do math (seriously, West Point required cadet take Calc 1, Calc 2, Prob and Stat, and a bunch of engineering classes...I totally got this). A single player in a 12v12 is 8.33% of the team. the other 91.7% is random. You cannot possibly think that over time an ~92% random event is going to produce a statistically meaningful result. And I haven't even started factoring in enemy pilot skill (random), enemy tonnages (random), enemy coordination (random), and so on. There. Is. Nothing. Statistical. About. Solo. Dropping.
Now, I'm absolutely capable of farming damage in a match even with potato teammates. That's the closest stat you can find that you can control, and even then you're likely only going to be able to push an extra 150-200 damage versus what you would have gotten fighting alongside your team. It's boring tho. Instead, I do silly things like try to sneak a stealth armor Griffin 2N up behind assaults and blow them to smithereens, and as you might have guessed, my results vary wildly and usually due to factors beyond my control...like getting spotted by a competent player who just happened to be facing my way when I tried to dash from one building to another. Or I run a narc Raven knowing I'm giving up pretty much any chance of a kill/good damage score/anything that shows up on the stat board. Or I try out a build that sounds good on paper to see if it actually works in reality (like stripping all the armor from a Mist Lynx in order to fit in 2xERLL and ECM...it works but your teammates will hate you).
So again, the leaderboard is garbage. Big, smelly, nasty, fly-infested with maggots overflowing garbage.
#14
Posted 02 November 2017 - 12:49 PM
So unless you seriously believe somehow 1) you're always put on the "bad team" and the enemy is always the "good team", 2) somehow other players are always on the "good team" and their enemy is the "bad team", what you're saying makes no sense.
Every player in the game has the same opportunity (over the long run) to fail with bad teams, carry mediocore teams, get carried by good teams in solo queue. The better players have a wider range of circumstances in which they can succeed and if we're talking only about solo queue, is generally reflected in stats (though leaderboards don't tell the whole story)
Edited by Mighty Wings, 02 November 2017 - 12:55 PM.
#15
Posted 02 November 2017 - 12:52 PM
Xavori, on 02 November 2017 - 12:19 PM, said:
No.
There is no amount of skill that can make up for a team that doesn't focus fire, scatters, and can't hit the broad side of the barn. Perhaps if we drop to 8v8 one player might be enough to tip the odds, but again, at only 12.5% of the team, it's very iffy.
And statistics blah blah blah. You want math, I can totally do math (seriously, West Point required cadet take Calc 1, Calc 2, Prob and Stat, and a bunch of engineering classes...I totally got this). A single player in a 12v12 is 8.33% of the team. the other 91.7% is random. You cannot possibly think that over time an ~92% random event is going to produce a statistically meaningful result. And I haven't even started factoring in enemy pilot skill (random), enemy tonnages (random), enemy coordination (random), and so on. There. Is. Nothing. Statistical. About. Solo. Dropping.
Now, I'm absolutely capable of farming damage in a match even with potato teammates. That's the closest stat you can find that you can control, and even then you're likely only going to be able to push an extra 150-200 damage versus what you would have gotten fighting alongside your team. It's boring tho. Instead, I do silly things like try to sneak a stealth armor Griffin 2N up behind assaults and blow them to smithereens, and as you might have guessed, my results vary wildly and usually due to factors beyond my control...like getting spotted by a competent player who just happened to be facing my way when I tried to dash from one building to another. Or I run a narc Raven knowing I'm giving up pretty much any chance of a kill/good damage score/anything that shows up on the stat board. Or I try out a build that sounds good on paper to see if it actually works in reality (like stripping all the armor from a Mist Lynx in order to fit in 2xERLL and ECM...it works but your teammates will hate you).
So again, the leaderboard is garbage. Big, smelly, nasty, fly-infested with maggots overflowing garbage.
What you are conveniently overlooking is that the random variables apply to everyone equally, especially over large sample sizes. The leaderboard itself demonstrates the fallacy of your argument, Xavori. If individual skill was immeasurable because of the randomness, there would be virtually no statistical deviation across the board. Everyone would end up with a 1.0 K/D and a 1.0 W/L within a small margin of error. The very fact that the best players repeatedly end up at the top of the boards, the potatoes at the bottom and the average players in the middle indicates that it is functioning at least marginally well. While it is not perfect, there is no arguing that it does not move players up and down according to performance. That is why most people who use it for reference also look at Average Match Score, because it most accurately reflects a players overall contribution to winning since it looks at total score and not just kills. This includes things like lance in formation, capping, spotting, and other non damage factors as well as damage, kills, etc... This allows "specialized" mechs and pilots to have some measure of contribution if damage is not their primary role.
The only reasonable argument to be made that the leaderboard doesn't reflect individual pilot skill is that organized team play will positively impact some of the personal stats where players can get carried to some extent, and in other cases may be carrying a team. But again, over a large sample size individual match anomalies are leveled out and a reasonable picture of individual skill emerges.
It is also observable if you take the time to watch the stats for yourself and your teammates. I've watched a dozen or more people in our unit whose position on the leaderboard rises steadily and reflects what is readily observable in their play in game as well.
#16
Posted 02 November 2017 - 01:23 PM
The Cyberserker, on 02 November 2017 - 12:17 PM, said:
Not true.My current K/D is 0.7 since I've been playing solo in mostly lights since I got back to the game. Before I quit it was getting close to 3 because I played almost exclusively in organized groups and larger mechs. Stats are an accurate representation of some things, but individual skill isn't one of them.
Yes they are though. You can look at the stats between different weight classes, it might tell you that you're better in heavy mechs than you are in lights.
I agree that having solo and group stats combined is kind of dumb, but if you control for that you can still gather a lot of useful information. Even if you don't control for that you can gather useful things. Your MS will tend to go down if you play in groups. Looking at any one single stat isn't that helpful, but looking at them as an aggregate absolutely will help determine how good a pilot is.
Xavori, on 02 November 2017 - 12:19 PM, said:
No.
There is no amount of skill that can make up for a team that doesn't focus fire, scatters, and can't hit the broad side of the barn. Perhaps if we drop to 8v8 one player might be enough to tip the odds, but again, at only 12.5% of the team, it's very iffy.
And statistics blah blah blah. You want math, I can totally do math (seriously, West Point required cadet take Calc 1, Calc 2, Prob and Stat, and a bunch of engineering classes...I totally got this). A single player in a 12v12 is 8.33% of the team. the other 91.7% is random. You cannot possibly think that over time an ~92% random event is going to produce a statistically meaningful result. And I haven't even started factoring in enemy pilot skill (random), enemy tonnages (random), enemy coordination (random), and so on. There. Is. Nothing. Statistical. About. Solo. Dropping.
Now, I'm absolutely capable of farming damage in a match even with potato teammates. That's the closest stat you can find that you can control, and even then you're likely only going to be able to push an extra 150-200 damage versus what you would have gotten fighting alongside your team. It's boring tho. Instead, I do silly things like try to sneak a stealth armor Griffin 2N up behind assaults and blow them to smithereens, and as you might have guessed, my results vary wildly and usually due to factors beyond my control...like getting spotted by a competent player who just happened to be facing my way when I tried to dash from one building to another. Or I run a narc Raven knowing I'm giving up pretty much any chance of a kill/good damage score/anything that shows up on the stat board. Or I try out a build that sounds good on paper to see if it actually works in reality (like stripping all the armor from a Mist Lynx in order to fit in 2xERLL and ECM...it works but your teammates will hate you).
So again, the leaderboard is garbage. Big, smelly, nasty, fly-infested with maggots overflowing garbage.
I don't care what courses you have taken. You are demonstrating that you lack a fundamental understanding of basic statistics with your argument. Clearly you didn't learn too much in statistics at West Point or you would realize how ridiculous your argument is.
I'm not arguing that one player can carry every match, even the best players will absolutely have teams they can't carry. If both teams were infinitely large (and evenly distributed) everyone would trend towards a 1.0 WLR. On the other extreme with a team size of one the best player could potentially have a perfect WLR.
A single player is 8.33% of the players, but can contribute more than 8.33% of the damage/kills/effort. You are correct in that 11 of the players on your team are randomly assigned and you can't control them.
What you are neglecting is that every other player also has the same 91.7% random on their team, so on average the only difference between the teams you are on and the teams you play against is you vs the player replacing you on the other side. You and that player have an equal chance of being on either side. Over a large number of matches that difference will become statistically significant.
Put a different way, give 100 matches with two completely random teams 12v12 let's assume you can get the following results
12:0
12:1
12:3
12:4
12:5
12:6
12:7
12:8
12:9
12:10
12:11
and the reverse. I'm ignoring the timer running out and ties.
In a 12:0 stomp one player clearly isn't going to make a difference. In a 12:11 game it literally comes down to a 1v1 battle at the end of the game. I hope you would agree that in this case the skill of the last player on each team makes a significant difference. Similarly for close matches 12:10, 12:9, 12:8, a good player could make the difference in winning or losing the match. While in a 12:1, 12:2, 12:3 game there really isn't much they could do.
If for random teams there is some average chance of getting each of these results then if you randomly replace one of these players with yourself you will either increase or decrease the chance of the team winning in the very close matches.
Do you think that a single player can make the difference in a match that is 12:11? If you do then over time there will be x% of matches that will be swung from loss to win by that player replacing a worse player (or vice versa). The better the player the more matches they can swing. A great player for example might be able to take a match that would have be 6:12 and turn it into a 12:11 win.
If instead of a good player we put a bad player in they might make their team lose a match 11:12 instead of winning 12:11. Over a large number of matches this will have a statistical effect of bringing down the WLR of that player.
#17
Posted 02 November 2017 - 01:55 PM
Tralik, on 02 November 2017 - 12:52 PM, said:
What you are conveniently overlooking is that the random variables apply to everyone equally, especially over large sample sizes. The leaderboard itself demonstrates the fallacy of your argument, Xavori. If individual skill was immeasurable because of the randomness, there would be virtually no statistical deviation across the board. Everyone would end up with a 1.0 K/D and a 1.0 W/L within a small margin of error. The very fact that the best players repeatedly end up at the top of the boards, the potatoes at the bottom and the average players in the middle indicates that it is functioning at least marginally well. While it is not perfect, there is no arguing that it does not move players up and down according to performance. That is why most people who use it for reference also look at Average Match Score, because it most accurately reflects a players overall contribution to winning since it looks at total score and not just kills. This includes things like lance in formation, capping, spotting, and other non damage factors as well as damage, kills, etc... This allows "specialized" mechs and pilots to have some measure of contribution if damage is not their primary role.
The only reasonable argument to be made that the leaderboard doesn't reflect individual pilot skill is that organized team play will positively impact some of the personal stats where players can get carried to some extent, and in other cases may be carrying a team. But again, over a large sample size individual match anomalies are leveled out and a reasonable picture of individual skill emerges.
It is also observable if you take the time to watch the stats for yourself and your teammates. I've watched a dozen or more people in our unit whose position on the leaderboard rises steadily and reflects what is readily observable in their play in game as well.
If I were so inclined, I could absolutely push my ranking on the leaderboard. I'd quit running undergunned mechs. I'd quit going ninja. (except I'd so miss moments like this: https://steamuserima...BCBC49F86C0454/ ) I'd also prolly quit playing MWO.
I do agree that as a player improves in skill, they'll move up some on the leaderboard. But I could move up even farther on the leaderboard just by caring about the leaderboard (I never will) because it's bad. It tracks to things like kills which are heavily influence by luck as they are last hit. I tracks to damage, and I could also easily push my damage numbers up by running my Tempest lurm boat because it's really good at farming damage even if it's really bad for pretty much everything else. I could run my comp build mechs all the time and get a lot more KMDD's/kills/damage. But again...bored now.
I mean, I like taking something silly like turning a cyclops into a walking UAV with sensor detection just past 1500m:
https://steamuserima...4088DE5AC259B4/
And then doing this with it:
https://steamuserima...2C46E302975378/
And this:
https://steamuserima...B45F27823DC6B7/
Or looking at it another way...
In scouting matches where I'm 1/4th of the team, I can almost always be counted on to have the highest damage because I can flat out carry (well, until the IS got all Bushwacker happy) 3 potatoes when I'm trying to. Ya know, like three solo kills in a scouting match:
https://steamuserima...22DFE0C5EDC7F1/
I can even have the highest damage when I'm running with our unit..like far and away the most damage...
https://steamuserima...088C8A6B347698/
And I have tons and tons of those screenshots (cuz I'm saving them up to eventually get all the unit specialist tags we have *smirk*)
But I flat out don't care about my leaderboard ranking...and my ranking reflects that even tho when I do play comp-ish, I'm really good. And as long as the leaderboard is based on so very many random elements, it's meaningless outside of "Can you game the system to improve the ranking?"
And since the answer is yes, it's pretty meaningless.
#18
Posted 02 November 2017 - 02:08 PM
Xiphias, on 02 November 2017 - 01:23 PM, said:
Random numbers are still random.
There are too many variables that are absolutely beyond your control to take anything meaningful out of the leaderboards. If you get lucky and get good teammates that work together your stats will be better. If you get even luckier and get the last hits on components or mechs, your damage and potentially KMDD's go up. Your win/loss record will go up.
Conversely, if you get bad teams, I don't care if you're the greatest player to ever drive a big stompy robot. Your prolly going to lose the match. Your damage numbers will go down because you're going to get overrun before you can do a lot of damage.
"But over time..." blah blah blah.
Over time, random numbers are still random.
#19
Posted 02 November 2017 - 03:07 PM
Xavori, on 02 November 2017 - 01:55 PM, said:
No one is saying that you can't improve your stats by trying to improve your stats. That's like a person saying they could lose weight if they wanted to. Sure they could, but they would have to do things like eat better and exercise more. Your basically saying that if you focused on getting more kills and dying less you'd have a higher KDR. Almost like if you focused on winning you could win more.
Quote
https://steamuserima...4088DE5AC259B4/
*Runs LRMs on assault mechs. Check
Quote
https://steamuserima...22DFE0C5EDC7F1/
I can even have the highest damage when I'm running with our unit..like far and away the most damage...
https://steamuserima...088C8A6B347698/
That's not even efficient damage. You got 2 kills with 950 damage in scouting. 99% certainty that you are boating streaks
*Boats streaks in scouting and thinks that makes him good. Check
Yeah, I can carry too.
4 solo Clan
4 solo IS (this was 4v1 too, fun game, DS came down right on top of two of them and the other two tried to make a dive for it)
Quote
And since the answer is yes, it's pretty meaningless.
That's fair that you don't care about the rank. I'm fine with that, more power to you. You're basically saying that you are playing worse than you could. Your teams lose more than they win though, regardless of how good you could play, you are actively contributing to your team's chances of losing through how you actually play.
Look at the rosters of top teams like EMP or Div A in MRBC it's funny how the better players tend to have the better stats over a large number of matches. You can only game the system so much. If a player can't hit the broad side of a barn no amount of gaming the system will get them a MS average of 500.
Xavori, on 02 November 2017 - 02:08 PM, said:
There are too many variables that are absolutely beyond your control to take anything meaningful out of the leaderboards. If you get lucky and get good teammates that work together your stats will be better. If you get even luckier and get the last hits on components or mechs, your damage and potentially KMDD's go up. Your win/loss record will go up.
Conversely, if you get bad teams, I don't care if you're the greatest player to ever drive a big stompy robot. Your prolly going to lose the match. Your damage numbers will go down because you're going to get overrun before you can do a lot of damage.
"But over time..." blah blah blah.
Over time, random numbers are still random.
It's not really surprising that you think component destruction is random. You do have to get pretty luck when you're boating streaks. That said, if you run direct fire weapons and have good aim/timing you can absolutely increase the amount of components you destroy.
Do you really think players are getting lucky/unlucky for 1000's of matches in a row? Or that their luck just tends to be good or bad and no matter how many games they play they are stuck with their luck?
Let me try this one more time.
Who do you think has a better chance of winning when they play each other, a team with 11 players or a team with 12 players?
Clearly, the team with 12 players is more likely to win. If we randomly assign 1000 sets of matches 12 vs 11, with random teams each time, on average the teams with 12 players will have a higher WLR than the teams with 11 players.
Now let's extend this to 12v12. If I create an alt account and disconnect/afk every match it is effectively making the game a 12v11. In that case would you say that my actions contribute to my team losing more often on average and my WLR being lower?
If you accept that a player's actions can contribute to their team losing more on average the reverse also has to be true, that a player's actions can contribute to helping the teams they are on win more frequently, on average.
If a player's actions can contribute to their teams winning more on average, that means they can affect their WLR and that WLR indicates the effect they are having on their team's chances of winning on average.
Imagine our random players are dice. We have 23 6 sided dice and 1 20 sided die. We randomly group these dice into "teams" of 12 and roll 1000 "matches" where the value of all the dice on a team are added together and compared to the other team. The team with the highest total wins. We track each die and record whether or not it won the match that it was in. Over the 1000 matches the 20 sided die will have a higher WLR than the 6 sided dice because it is contributing to its team winning on average. If you through different types of dice into the mix, say a 4 sided die it doesn't chance the overall result. The 20 sided die will have a higher WLR because it is "better" and the 4 sided die will have a lower WLR because it is worse. The 20 sided die can't always win the match, but can be the deciding factor in closer matches because it can do the work of 3.3 6 sided dice or 5 4 sided dice. This is the same situation as MWO.
Random numbers are random, but a player's ability is not a random number. When combined into a group of random numbers it can still have a significant effect.
You literally don't understand how statistics work.
#20
Posted 02 November 2017 - 03:22 PM
1 user(s) are reading this topic
0 members, 1 guests, 0 anonymous users