Jump to content

Devs Come In. How Are These People Still In T1 Matches?


119 replies to this topic

#101 Xiphias

    Member

  • PipPipPipPipPipPipPip
  • Littlest Helper
  • Littlest Helper
  • 862 posts

Posted 14 December 2017 - 08:32 AM

View PostLily from animove, on 12 December 2017 - 06:26 AM, said:

sorry, but thats not the average numbers, because according to this it would need averagely 250-300 damage to destroy a mech and thats far above what really is needed.

1500 is too high, but 1000-1100 really is a solid average to expect for FP. Taking the matches that I have on imgur as a sample I get the following results for my teams:

1045 Average damage pug match (Personal 204 damage per kill)
1110 Average damage pug match (Personal 223 damage per kill)
1145 Average damage pug loss (Personal 267 damage per kill, inflated by loss)
1245 Average damage pug match (Personal 189 damage per kill)
1185 Average damage group match (Personal 213 damage per kill)

While you can kill mechs with less than 250 damage FW tends to lead to inefficient damage. Lots of side torsos getting destroyed and strikes/LRMs spreading damage to components like legs. A lot of times one player will be shooting a mech on one side/component and the other will be shooting a different side/component. This inflates the damage instead of just what is needed to kill a mech.

Sure, an elite team might be able to kill the other side with less damage, but most teams aren't going to be that competent. Also, if a player is that good they should be getting significantly more kills.

It's reasonable to expect players to get a minimum of 1000 damage to be contributing. If a player is getting less than 1000 damage they are being a drag on their team.

#102 Lily from animove

    Member

  • PipPipPipPipPipPipPipPipPipPipPip
  • The Devoted
  • The Devoted
  • 13,891 posts
  • LocationOn a dropship to Terra

Posted 14 December 2017 - 09:43 AM

View PostXiphias, on 14 December 2017 - 08:32 AM, said:

1500 is too high, but 1000-1100 really is a solid average to expect for FP. Taking the matches that I have on imgur as a sample I get the following results for my teams:

1045 Average damage pug match (Personal 204 damage per kill)
1110 Average damage pug match (Personal 223 damage per kill)
1145 Average damage pug loss (Personal 267 damage per kill, inflated by loss)
1245 Average damage pug match (Personal 189 damage per kill)
1185 Average damage group match (Personal 213 damage per kill)

While you can kill mechs with less than 250 damage FW tends to lead to inefficient damage. Lots of side torsos getting destroyed and strikes/LRMs spreading damage to components like legs. A lot of times one player will be shooting a mech on one side/component and the other will be shooting a different side/component. This inflates the damage instead of just what is needed to kill a mech.

Sure, an elite team might be able to kill the other side with less damage, but most teams aren't going to be that competent. Also, if a player is that good they should be getting significantly more kills.

It's reasonable to expect players to get a minimum of 1000 damage to be contributing. If a player is getting less than 1000 damage they are being a drag on their team.


these stats are somewhat better yetl quite invalid because they are biased towards your team you seem to drop as a group and that group existing of players above averages, or are biased towards the team the opponent dropped as. They aren't true pug matches because this would mean, both sides are pugs. That kinda makes the average of your matches quite different and not a valid unbiased selection of true skill distributed performers..

further all your games were focussed on killing the opponents, How do you handle groups that focus so hard on objectives that they don't een kill many mechs? they will average get a lot less score since score is mostly based on damage.

further, you are ENTIRELY doing statistics wrong. You make a arithmetical average, but this is not how you average the perfoemance of the average pilot.

First every scratched mech adds damage towards your damage/mech calculation while this damage isn't the one that was killing as well.
second When a good player (which often hapens) is on the highly winning side they tend to farm some damage. This invalides true numbers, and even mroe worse it increases the damage/kill but not for the average pilot, only for the above average pilot. You should go and look how these numebrs alter when you play durign a FW event when all

the pugs come to the field as well, and then go and play on a pug and not a preset team. Things then are entirely different.

take these 2 screens

https://i.imgur.com/I8yDxRs.jpg
https://i.imgur.com/l2VMYBt.jpg

entiely different running matches. But they show you how averages as youc alculate them do not work.

I have a rather thrown together bakcup fo some score screens here https://imgur.com/a/VNVw7
When you browse through them, youw ill see some pilots do 10x as much score and damage than others. Those are the reason why an arithmetical average is absolutely not the right way to judge the average pilot's performance. because a pilot 10x as good as the worst doesn't means the average is at 5x as good as the worst. This si simply by how skills distribut ein MWO not a working method.

The averages players performance is where 50% of the population performs, it is NOT where the average Score/damage is. because to make this happen, player skill would require a perfect bell distribution around the avergage, but MWo palyerbase is not distributed liek that. They are a very hunchy bell with a LARGE "lowie" side and a small high side.

So when we speak a about skill distibution of palyers we need to talk about the median not the average performance. the averages pilots performance is basically the performance of the medians pilots performance. thats because this is where half the players are above (performance wise) and the other half of the players are beyond (performance wise)

The simplified example why your's isn't working is:

pilot one makes 2 kills with 800damage, and pilot 2 and 3 do 150damage and 0 kills and pilot 4 one kill with 50damage.

the reuslt is:

1100 damage for 3 kills averaging 333damage/kill.

but the average pilots done damage IS NOT 333 damage. it is more around 150 damage. Even 1100 combiend damage divided by 4 isnt going to be the required 275 as mathermatical average, thats still nearly twice as much as that what 2 out of 4 pilots performed. And further it is far beyond what 75% of the pilots in thuis example are even able to perform.

That is why tarogatos excel is very good in explaining how performance of players truly distributes.

please again go into his excel

https://docs.google....#gid=1371966739

look at the match score sheet, it shoes a VERY important information about the raw numbers where borders for match score are based uppon population %

lets have a look at the non mech class based table.
50% of the people are doinng 191 score or LESS. so the average players performance is 191 score or less. The absolute average is 201 Score, thats quite a difference when you look how the % change over scores. 25% of the ones above the 50% population only make 33score more. or with other words, 75% of the pilots score less than 224 score.

Ok, this si lookign at seasn 3, you can nicely see when some wepaons got nerfed and stuff with tech happened, as it shifted numbers. But funnily the 50% of S1 and S17 vs the 25% or 75% have a quite similar delta to each other.

But summing up damage/scores and dividing it by pilots or killed mechs is definately not giving your the correct average pilots average performance. Esepcially nto when you take matches where YOU are included or anyone else that performans like 5x as much as a scrub. because then you are statistically ruining the this specifics matches true average perormance upwards for the average pilot.

I hope you understand what I see, and can see why the kind of calculation you do is not giving us areal average for the average population.

yes you can expect the current FP players to have at leats 1k damage, but this isn't meaning the average pilots performance, because a lot lower skilled players don't even play FP. But usually if I go through my screens, the 1kdamage is often met by 7-13 people as far as i saw quickly running over my screens. But if you head over to the Solo Q now, and would have no MM, you could expect the true pilots average to be a lot below 250 damage.because in that environment a lot more low skille dpalyers who never even try to touch FP are involved as well. So my guess for the current FP is that the average pilots true damage is somewhere between 170 and 200. whcih means everoyne with like 800 damage did at leats carry his own weight properly. all the others below did rather poorly.

Edited by Lily from animove, 14 December 2017 - 09:51 AM.


#103 Xiphias

    Member

  • PipPipPipPipPipPipPip
  • Littlest Helper
  • Littlest Helper
  • 862 posts

Posted 14 December 2017 - 12:46 PM

View PostLily from animove, on 14 December 2017 - 09:43 AM, said:

these stats are somewhat better yetl quite invalid because they are biased towards your team you seem to drop as a group and that group existing of players above averages, or are biased towards the team the opponent dropped as. They aren't true pug matches because this would mean, both sides are pugs. That kinda makes the average of your matches quite different and not a valid unbiased selection of true skill distributed performers..

Only one of those matches was I in a group. I think another one had a decent size group on my side. The others might have had some small groups, but for practical purposes were pug groups. They are fairly representative of FP.

I have plenty more screenshots, however they aren't uploaded and I'm not at my desktop at the moment. I have the feeling they would support the trend.

Quote

further all your games were focussed on killing the opponents, How do you handle groups that focus so hard on objectives that they don't een kill many mechs? they will average get a lot less score since score is mostly based on damage.

I think rushing objectives in FP is dumb and boring. Sure those players will average less damage, what of it? The issue at hand is that in a fight you should be able to expect an average/competent player to get 1000 damage. I'm not looking at matches like that because obviously that is going to reduce the damage.

Quote

further, you are ENTIRELY doing statistics wrong. You make a arithmetical average, but this is not how you average the perfoemance of the average pilot.

I'm not doing statistics at all. You stated that there wasn't enough damage for each player to do 1000 damage, I posted the screenshots I had available to show that that isn't true. There is enough damage on a winning team (or even a close loss) for 12 players to all get 1000 damage. I'm not saying they all will, but that the damage is there.

Some players are going to score more than the average damage, some are going to score less. If you're scoring less than the average you're by definition below average at scoring damage. It's not a rigorous statistical analysis, it's a back of the envelope calculation that shows it's possible for 12 players to all get 1000 or more damage.

Quote

First every scratched mech adds damage towards your damage/mech calculation while this damage isn't the one that was killing as well.
second When a good player (which often hapens) is on the highly winning side they tend to farm some damage. This invalides true numbers, and even mroe worse it increases the damage/kill but not for the average pilot, only for the above average pilot. You should go and look how these numebrs alter when you play durign a FW event when all

the pugs come to the field as well, and then go and play on a pug and not a preset team. Things then are entirely different.

Scratch damage is still damage. We aren't talking about efficient damage here. A player who can do efficient damage is going to have no problem getting above 1k anyway.

Quote

take these 2 screens

https://i.imgur.com/I8yDxRs.jpg
https://i.imgur.com/l2VMYBt.jpg

entiely different running matches. But they show you how averages as youc alculate them do not work.

The winning teams in those games have averages of 1072 damage and 988 damage respectively, which reinforces my point that an average damage of 1000 is a fair range to expect.

I'm not sure what your problem with the calculation is.

Quote

I have a rather thrown together bakcup fo some score screens here https://imgur.com/a/VNVw7
When you browse through them, youw ill see some pilots do 10x as much score and damage than others. Those are the reason why an arithmetical average is absolutely not the right way to judge the average pilot's performance. because a pilot 10x as good as the worst doesn't means the average is at 5x as good as the worst. This si simply by how skills distribut ein MWO not a working method.

The averages players performance is where 50% of the population performs, it is NOT where the average Score/damage is. because to make this happen, player skill would require a perfect bell distribution around the avergage, but MWo palyerbase is not distributed liek that. They are a very hunchy bell with a LARGE "lowie" side and a small high side.

Sorry, where is your evidence for this?

Using the global stats from season 16 and using match score as a proxy for damage I get this:
Posted Image

Which looks like a fairly normal distribution to me.

Quote

Lots of stuff explaining stats

I understand all this and how stats work.

I should have been more precise in my phrasing. For a winning team it's perfectly reasonable to expect the average damage to be at least 1000 damage if not more (the loss I posted has the winning team at 1605). I've been in plenty of matches where the whole team scored above 1000 (obviously not typical of pugs). The point being for FP in particular, being competent means doing more than 1k damage.

Quote

I hope you understand what I see, and can see why the kind of calculation you do is not giving us areal average for the average population.

yes you can expect the current FP players to have at leats 1k damage, but this isn't meaning the average pilots performance, because a lot lower skilled players don't even play FP.

And my point is that you should expect the current FP players to have at least 1k damage and that if players can't play at that level they probably aren't ready for FP. Again, it was a back of the envelope calculation, not a rigorous statistical analysis.

Quote

But usually if I go through my screens, the 1kdamage is often met by 7-13 people as far as i saw quickly running over my screens. But if you head over to the Solo Q now, and would have no MM, you could expect the true pilots average to be a lot below 250 damage.because in that environment a lot more low skille dpalyers who never even try to touch FP are involved as well. So my guess for the current FP is that the average pilots true damage is somewhere between 170 and 200. whcih means everoyne with like 800 damage did at leats carry his own weight properly. all the others below did rather poorly.

I don't really care what the average QP player does in a match. What's relevant is what the average FP player does in a match, and more importantly what the average winning FP player does in a match.

While 800 might be the true average, the winning average is going to be closer to or above 1000 because players that are consistently scoring below 1000 are being a drag on their team. Obviously the exact number doesn't have to be 1000, it could be higher or lower, but it tends to be pretty accurate in my experience.

Let me put it this way, pretend it take 12,000 damage to kill the entire enemy team and win the match (I think this is reasonable based on the screenshots that I posted). That means that with a perfectly even team you would expect each player to score 1000 damage. If a player scores less that 1000 damage they can be considered a negative contribution and if they score more than 1000 damage they can be considered a positive contribution.

If a player is consistently scoring below 1000 damage in FP they are negatively contributing to their team and the further below 1000 the more negative the contribution. I don't really care if instead of a bell curve we have a double hump and the average damage score is 600. That just means that the average player is having a negative contribution to winning.

#104 Kin3ticX

    Member

  • PipPipPipPipPipPipPipPipPip
  • The People's Hero
  • The People
  • 2,926 posts
  • LocationSalt Mines of Puglandia

Posted 14 December 2017 - 12:52 PM

View PostKoniving, on 08 December 2017 - 05:00 PM, said:

Easy.

Stats.
Are,.
MEANINGLESS (in MWO)

From another thread.


lol no

#105 Koniving

    Welcoming Committee

  • PipPipPipPipPipPipPipPipPipPipPipPipPip
  • The Guide
  • The Guide
  • 23,384 posts

Posted 14 December 2017 - 01:16 PM

View PostKin3ticX, on 14 December 2017 - 12:52 PM, said:


lol no

Well, there's a world class top tier guy who died with only 12 damage.

Doesn't mean he is terrible, after all he is world class top tier and a competition player.

But clearly from that one stat, taken out of context, he's on par with the worst players in the game.

Since our stats are NOT recorded accurately, ALL of our stats are Out of Context... and untrue.
Therefore all assertions derived from such facts are as true as this image.
Posted Image
But hey, accurate stats, like accurate image comparisons, aren't needed in court of the foolish.

....Doesn't mean they are any less meaningless just because the ignorant worship such inaccurate numbers as idols depicting how great they are, so "lol yes."

Edited by Koniving, 14 December 2017 - 01:18 PM.


#106 Kin3ticX

    Member

  • PipPipPipPipPipPipPipPipPip
  • The People's Hero
  • The People
  • 2,926 posts
  • LocationSalt Mines of Puglandia

Posted 14 December 2017 - 01:19 PM

View PostKoniving, on 14 December 2017 - 01:16 PM, said:

Well, there's a world class top tier guy who died with only 12 damage.

Doesn't mean he is terrible, after all he is world class top tier and a competition player.

But clearly from that one stat, taken out of context, he's on par with the worst players in the game.

Since our stats are NOT recorded accurately, ALL of our stats are Out of Context... and untrue.
Therefore all assertions derived from such facts are as true as this image.
Posted Image
But hey, accurate stats, like accurate image comparisons, are holdable in court.

....They are still just as meaningless, so "lol yes."


Well if you take business statistics or econometrics you will learn you dont just go off of one 12 damage game and declare stats meaningless.

In the long run, if your stats are garbo, you cant explain it away.

Everything you say about stats in mwo being broken is bunk. Just stop, plz. Go find something else to be wrong about.

Edited by Kin3ticX, 14 December 2017 - 01:22 PM.


#107 Xiphias

    Member

  • PipPipPipPipPipPipPip
  • Littlest Helper
  • Littlest Helper
  • 862 posts

Posted 14 December 2017 - 03:47 PM

View PostKoniving, on 14 December 2017 - 01:16 PM, said:

Well, there's a world class top tier guy who died with only 12 damage.

Doesn't mean he is terrible, after all he is world class top tier and a competition player.

Outliers don't disprove statistical trends.

Let's look at his stats, unsurprisingly he's listed in the top 1%. Now let's look at the stats of all the finalist teams:

EMP
EON
228

All of the players on those lists are in the top 3% and most of them are in the top 1% or higher. Coincidence?

Quote

But clearly from that one stat, taken out of context, he's on par with the worst players in the game.

Since our stats are NOT recorded accurately, ALL of our stats are Out of Context... and untrue.
Therefore all assertions derived from such facts are as true as this image.

No one is claiming that the results of a single match are indicative of a player's skill. Everyone has bad matches.

The stats that we have while not perfect are sufficient for making some broad generalizations. There is enough context to tell that a person with a thousand matches and a WLR of 0.5 is not a world class player. You can absolutely draw conclusions with a pretty high certainty provided sufficient matches have been played and that there is a reasonable difference between player scores. You can't say that number 10 on the Jarl's list is better than number 11, but you can be pretty confident that number 1000 is better than number 10,000.

Admittedly we don't have perfect data, but even with flawed data it's still possible to draw conclusions with a fairly high certainty. It won't always be right, but if stats let me pick the better player 90% of the time that's still pretty good and a useful tool.

#108 sub2000

    Member

  • PipPipPipPipPip
  • Bad Company
  • Bad Company
  • 127 posts

Posted 14 December 2017 - 03:59 PM

View PostLily from animove, on 14 December 2017 - 09:43 AM, said:

a lot of not necessary words

1*..1500 comes from "requirement" to carry pug team and obviously has nothing to do with "average" or what not.The argument is "if you want to enjoy pugging in FW you need to be able to do 1500damage consistently". If to bother it's not difficult to actually dig direct quotes coming from different people at different times in FW threads. Look yourself.
2*.. Average has nothing to do with FW. Period. The whole direction of your arguments is beyond strange.
General request is not to be a potato and be able to dish 1000damage (250x4mechs). To kill typical IS mech one need 200damage to drill CT, usual damage to kill a moving active mech is close to 300 and with erlarge, lrms, srms etc. the number easily reaches 400 per mech. Or more.

#109 Koniving

    Welcoming Committee

  • PipPipPipPipPipPipPipPipPipPipPipPipPip
  • The Guide
  • The Guide
  • 23,384 posts

Posted 14 December 2017 - 03:59 PM

View PostXiphias, on 14 December 2017 - 03:47 PM, said:

Outliers don't disprove statistical trends.

Let's look at his stats, unsurprisingly he's listed in the top 1%. Now let's look at the stats of all the finalist teams:

EMP
EON
228

All of the players on those lists are in the top 3% and most of them are in the top 1% or higher. Coincidence?


No one is claiming that the results of a single match are indicative of a player's skill. Everyone has bad matches.

The stats that we have while not perfect are sufficient for making some broad generalizations. There is enough context to tell that a person with a thousand matches and a WLR of 0.5 is not a world class player. You can absolutely draw conclusions with a pretty high certainty provided sufficient matches have been played and that there is a reasonable difference between player scores. You can't say that number 10 on the Jarl's list is better than number 11, but you can be pretty confident that number 1000 is better than number 10,000.

Admittedly we don't have perfect data, but even with flawed data it's still possible to draw conclusions with a fairly high certainty. It won't always be right, but if stats let me pick the better player 90% of the time that's still pretty good and a useful tool.

You're entering the conversation rather late. It has been proven that the stats provided by PGI are incomplete and do not add up, even when you account for how very strangely they are divided up into different places for reasons beyond comprehension.
When I originally posted a response here I had 371 matches unaccounted for in terms of stats. I also had more deaths than matches, despite not having participated in faction play for a few seasons. That is not possible, unless the stats are false or incomplete.

False stats or incomplete stats are stats out of context with no grounds for comparison and therefore in either case are lies. If we say a lie is true, we can determine anything.

For example I can prove I am the pope.
The pope is a man.
I am a man.
Therefore I am the Pope.

Doesn't make it true, but if that is all you have to go by then we can say anything we want.
Without all the facts, we can say that I am the pope because I am a man, and therefore fit the criteria of being the pope.

The stats are untrue, and therefore anything derived from them, regardless of how logical or illogical, is the same as using them out of context. We can make the assertion that X is good in general and it would be generally true, like the pope being a man. But to say I am the pope is a falsehood that is just a true in that regard. By that extension, anything we say by the stats are just as true as they are false.

Edited by Koniving, 14 December 2017 - 04:05 PM.


#110 Kin3ticX

    Member

  • PipPipPipPipPipPipPipPipPip
  • The People's Hero
  • The People
  • 2,926 posts
  • LocationSalt Mines of Puglandia

Posted 14 December 2017 - 04:12 PM

View PostKoniving, on 14 December 2017 - 03:59 PM, said:

You're entering the conversation rather late. It has been proven that the stats provided by PGI are incomplete and do not add up, even when you account for how very strangely they are divided up into different places for reasons beyond comprehension.
When I originally posted a response here I had 371 matches unaccounted for in terms of stats. I also had more deaths than matches, despite not having participated in faction play for a few seasons. That is not possible, unless the stats are false or incomplete.

False stats or incomplete stats are stats out of context with no grounds for comparison and therefore in either case are lies. If we say a lie is true, we can determine anything.

For example I can prove I am the pope.
The pope is a man.
I am a man.
Therefore I am the Pope.

Doesn't make it true, but if that is all you have to go by then we can say anything we want.
Without all the facts, we can say that I am the pope because I am a man, and therefore fit the criteria of being the pope.

The stats are untrue, and therefore anything derived from them, regardless of how logical or illogical, is the same as using them out of context. We can make the assertion that X is good in general and it would be generally true, like the pope being a man. But to say I am the pope is a falsehood that is just a true in that regard. By that extension, anything we say by the stats are just as true as they are false.


this is all nonsense

Come back when you get a 1.0 WL. Apparently nobody else can do it because they are getting crushed by PGI's broken stats (they can).

#111 Xiphias

    Member

  • PipPipPipPipPipPipPip
  • Littlest Helper
  • Littlest Helper
  • 862 posts

Posted 15 December 2017 - 01:18 AM

View PostKoniving, on 14 December 2017 - 03:59 PM, said:

You're entering the conversation rather late. It has been proven that the stats provided by PGI are incomplete and do not add up, even when you account for how very strangely they are divided up into different places for reasons beyond comprehension.
When I originally posted a response here I had 371 matches unaccounted for in terms of stats. I also had more deaths than matches, despite not having participated in faction play for a few seasons. That is not possible, unless the stats are false or incomplete.

I can't speak to exact numbers on stats, but I accept that there are some problems with them. That said, say 10% of a person's stats are incorrect (not just missing, but skewed in the same direction) that would indicate that they could vary about 20% (either up or down). You could still tell if a person had significantly better stats, even if 10% of each person's stats were wrong.

Quote

False stats or incomplete stats are stats out of context with no grounds for comparison and therefore in either case are lies.

Incomplete stats aren't a problem with a large sample size as long as they are randomly occurring. I rather doubt that it's only the top 10% of you matches that are disappearing. I also haven't seen any evidence of false stats significantly changing a player's stats.

I watch my stats and they tend to follow my performance pretty well. You're going to have to prove not just that there are some inaccuracies in the stats, but that the inaccuracies have a significant effect on a person's overall stats.

Quote

If we say a lie is true, we can determine anything.

For example I can prove I am the pope.
The pope is a man.
I am a man.
Therefore I am the Pope.

Doesn't make it true, but if that is all you have to go by then we can say anything we want.
Without all the facts, we can say that I am the pope because I am a man, and therefore fit the criteria of being the pope.

You're confusing absolute statements with probable statements. No statistic says that player x will always beat player y, just that player x has a 99% change of beating player y. It's not saying a lie is true, it's accepting that some uncertainty exists and concluding what the most probable outcome is.

To use your example, the pope is a man, you are a man, therefor you could be the pope. However, based on probability there are ~3.5 billion men in the world and only one pope. Therefore if a man claims to be the pope (lacking any additional evidence) it is always in my best interest to believe that they are not the pope, there is a 2.9*10^-8% chance it actually is the pope and a 99.999999971% chance that it isn't the pope.

Quote

The stats are untrue, and therefore anything derived from them, regardless of how logical or illogical, is the same as using them out of context. We can make the assertion that X is good in general and it would be generally true, like the pope being a man. But to say I am the pope is a falsehood that is just a true in that regard. By that extension, anything we say by the stats are just as true as they are false.

You don't need 100% accuracy to derive a logical conclusion. Imagine a counter case. I have a thermometer, the thermometer is always off by a degree, either up or down, I don't know which way, but I know that it is never correct. It is false, a lie if you will. Now imagine that I read an outside temperature of 100F. I draw the conclusion that it is hot. Is my conclusion false? Given the variance it could be 99F or 101F, either way it's still hot. Even if the thermometer has an error of 10 degrees (10%) it's still going to be between 90-110F, both still "hot" temperatures.

I can never say what the temperature is, but I can say whether or not it is hot or cold. Similarly, with stats I can't say how good a person is, but I can have a pretty good guess as to whether they are good or bad.

Even if there is some error in a person's statistics that doesn't mean their statistics can't be used to draw perfectly valid conclusions from. As long as the error is reasonably small and the sample size is reasonably large you can guess with reasonable confidence. You don't know, but you can be confident that it is likely.

#112 Dr Cara Carcass

    Member

  • PipPipPipPipPipPipPip
  • Guillotine
  • Guillotine
  • 643 posts

Posted 15 December 2017 - 03:20 AM

View PostTroa Barton, on 08 December 2017 - 05:05 PM, said:

Sorry Ive been watching too much sargon, by SJW do you mean social justice warrior or something else?
Because if you're accusing me of being one you're barking up the wrong tree.
I'm just here for the outdated dank memes.


Carefull with these names - MTG is banning pro players for openly admitting that they like someone who did throw some memes around twitter, or stated that cosplayer (meaning women in the MTG cossplay community) live of money from 12-14 year olds who throw daddys money after some half way visible ****.
Immagine that - you are apro player and are not allowed to play anymore because on some other platform yous aid that guy actually is right....

#113 Dr Cara Carcass

    Member

  • PipPipPipPipPipPipPip
  • Guillotine
  • Guillotine
  • 643 posts

Posted 15 December 2017 - 04:04 AM

View PostDakota1000, on 11 December 2017 - 09:40 AM, said:


Recent events show that Clan is winning during most of the event, but as soon as IS groups up on the last day they can just easily reverse the whole trend and bring it to a tie. The "if you want to win, group up" line still holds true. What's really making Clan pugs win more than IS pugs is just the level of customization available in IS mechs compared to Clan ones, bads can make even worse builds in IS mechs than they can with Clan ones.



I had a lot of free wins lately on IS side.
Like 4-5 In a row in invasion because nobody is playing FW anymore. If thats teh case for the rest yes you win because you get free drops - does not say anyhtign about balance....

#114 Lily from animove

    Member

  • PipPipPipPipPipPipPipPipPipPipPip
  • The Devoted
  • The Devoted
  • 13,891 posts
  • LocationOn a dropship to Terra

Posted 15 December 2017 - 04:22 AM

Xiophias, where are the numbers for that graph you posted? without that numbers I cnnnot explain you that as the numbers I would drag form it would be pretty rough estimations based on the grphics.

#115 The Lighthouse

    Member

  • PipPipPipPipPipPipPipPip
  • Moderate Giver
  • Moderate Giver
  • 1,142 posts

Posted 15 December 2017 - 04:58 AM

View PostKin3ticX, on 14 December 2017 - 04:12 PM, said:


this is all nonsense

Come back when you get a 1.0 WL. Apparently nobody else can do it because they are getting crushed by PGI's broken stats (they can).


Unfortunately Koniving is right about stats are broken. I am actually checking my mech stats and some of them are just completely wrong..... like stats showing that I played just couple of matches on my some of Javelins... despite I played them a lot enough to fully mastered them.

I don't think stats from PGI are that robust. I mean there was a case where Tarogato found his data and the data from Jarl's leaderboard have disagreements, as well as disagreements between the stat on my profile and Jarl's.







Edit : I am still checking my mech stats, and something is definitely very wrong. My Reaver match number is only 14, and according to the stat I only played my Invasion variant Shadow Cat just once.



Just once.



That mech even has historical xp points.



I don't know guys. MAYBE PGI-published stat can be still useful to a degree, if you played a lot of games. But you know, is a bit absurd that we are so dependent on a thing that is not even remotely robust in the first place, right?


...To be honest, the fact that I just found out surely shows I give zero damn about my own stat, but I do believe this should be concerning for those who play this game competitively.

Edited by The Lighthouse, 15 December 2017 - 05:27 AM.


#116 Wolfways

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Mercenary
  • 6,499 posts
  • LocationIn a shutdown overheated mech near you.

Posted 15 December 2017 - 05:38 AM

I'm in T2 and I'm still constantly in matches with players who:
Don't shoot UAV's down, or can't because they removed arm weapons,
Stand directly behind me when I'm trying to back up,
Run straight into my legs at the start of every match,
Half way through the match are running around in a Warhammer with no weapons because they only put 2xLBX in the st's and they are out of ammo,
...the list goes on.

I'm getting sick of playing again tbh Posted Image

#117 Asym

    Member

  • PipPipPipPipPipPipPipPipPip
  • Nova Captain
  • 2,186 posts

Posted 15 December 2017 - 10:34 AM

Some of you are really trying to "make a difference" and I applaud your efforts..... Unfortunately, as many have speculated, a lot of the data isn't valid.... It is impossible to make sense of anything with invalid data coming from a source that tends to have large amounts of variation.....

If I were to guess, since I don't see any PGI supplied and independently validated data, that low population and vast amounts of player variation are the culprits as to why there are so many player inconsistencies in every match we draw and play...

Neat discussion though !

#118 Xiphias

    Member

  • PipPipPipPipPipPipPip
  • Littlest Helper
  • Littlest Helper
  • 862 posts

Posted 15 December 2017 - 01:06 PM

View PostLily from animove, on 15 December 2017 - 04:22 AM, said:

Xiophias, where are the numbers for that graph you posted? without that numbers I cnnnot explain you that as the numbers I would drag form it would be pretty rough estimations based on the grphics.

You can find all the stats scraped from the leaderboards here:
https://drive.google...Y3VhVXdfcXE5a2c
https://www.reddit.c...aderboard_data/

These are the same stats that Tarogato's sheets are based off of.

#119 Xiphias

    Member

  • PipPipPipPipPipPipPip
  • Littlest Helper
  • Littlest Helper
  • 862 posts

Posted 15 December 2017 - 01:11 PM

View PostAsym, on 15 December 2017 - 10:34 AM, said:

Some of you are really trying to "make a difference" and I applaud your efforts..... Unfortunately, as many have speculated, a lot of the data isn't valid.... It is impossible to make sense of anything with invalid data coming from a source that tends to have large amounts of variation.....

Prove it.

You can absolutely draw conclusions from data that has large variations and no one has shown that the any problems with the stats have had a significant effect on a persons overall stats. The only thing I've heard is hand waving excuses that the "stats aren't valid".

I can play matches and show directly how they effect my stats. Can you demonstrate a significant divergence in a player's stats that's causing them to move up or down significantly?

#120 Suko

    Member

  • PipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 1,226 posts
  • LocationPacific Northwest

Posted 15 December 2017 - 01:31 PM

View PostBush Hopper, on 14 December 2017 - 07:32 AM, said:


I don't consider myself T1 material. Maybe T2. However, you shouldn't blame the players. They just play in the environment set up by PGI.



As someone already said: the problem is the PSR. You climb steadily. I think no one can be that bad as to lose ranks. This means, in theory, that ALL are T1 in the end.



This. A million times this. I am not T1. I'm T2 at best, but I've been sitting solidy in T1 since a month after PSR came out. What is really a shame is that the PSR is the same for ALL weight calsses and doesn't take long periods off from MWO into account.

I mostly play Heavy and Assault mechs. I am terrible in light mechs...like horrible. But because MWO thinks I am T1, no matter what mech I play in, I find it even more difficult to convince myself to try out lights (and sometimes mediums) because if I do I know I'll be hurting my team significantly. I want to get better at lights, but that's difficult when the game says I'm T1, but really I'm probably around T4 in that class.





1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users