Jump to content

Hammer's 3050 Challange Results (Lots Of Data Inside!)


89 replies to this topic

Poll: Mech diversity! (146 member(s) have cast votes)

Would you like to see more data like this in the future

  1. Yes (133 votes [91.10%])

    Percentage of vote: 91.10%

  2. No (5 votes [3.42%])

    Percentage of vote: 3.42%

  3. I don't care (8 votes [5.48%])

    Percentage of vote: 5.48%

Do you find the results interesting?

  1. What I expected (21 votes [14.38%])

    Percentage of vote: 14.38%

  2. Interesting (102 votes [69.86%])

    Percentage of vote: 69.86%

  3. Surprised (12 votes [8.22%])

    Percentage of vote: 8.22%

  4. Unimpressed (6 votes [4.11%])

    Percentage of vote: 4.11%

  5. Don't care. (5 votes [3.42%])

    Percentage of vote: 3.42%

Did this study change your stance/outlook on ECM/matchmaker

  1. Yes to both (17 votes [11.64%])

    Percentage of vote: 11.64%

  2. Yes to ECM (3 votes [2.05%])

    Percentage of vote: 2.05%

  3. Yes to Matchmaker (24 votes [16.44%])

    Percentage of vote: 16.44%

  4. No to either (89 votes [60.96%])

    Percentage of vote: 60.96%

  5. Don't care (13 votes [8.90%])

    Percentage of vote: 8.90%

Same as 3, but on mech diversity overall

  1. Yes (35 votes [23.97%])

    Percentage of vote: 23.97%

  2. No (67 votes [45.89%])

    Percentage of vote: 45.89%

  3. Maybe (33 votes [22.60%])

    Percentage of vote: 22.60%

  4. Don't care (11 votes [7.53%])

    Percentage of vote: 7.53%

When would you like to see the next diversity study done?

  1. During a tournament (4 votes [10.00%])

    Percentage of vote: 10.00%

  2. During a class weekend (0 votes [0.00%])

    Percentage of vote: 0.00%

  3. The week highlander is release (1 votes [2.50%])

    Percentage of vote: 2.50%

  4. The week after highlander release (6 votes [15.00%])

    Percentage of vote: 15.00%

  5. Monthly (23 votes [57.50%])

    Percentage of vote: 57.50%

  6. Other (please elaborate in comments) (6 votes [15.00%])

    Percentage of vote: 15.00%

Vote Guests cannot vote

#61 hammerreborn

    Member

  • PipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 3,063 posts
  • LocationAlexandria, VA

Posted 12 April 2013 - 06:04 AM

View PostKageru Ikazuchi, on 11 April 2013 - 08:14 PM, said:

Interesting ... doesn't jibe with my experience (pure PUG), but interesting.

If the long term statistics show that an ECM imbalance does not significantly affect the outcome of a match, I'll happily shut up (and maybe change my opinion) ... but when considering only matches where there was a significant ECM imbalance, I'm not convinced the sample size (36) is large enough to come to a valid conclusion, especially when the overall ECM mech population was lower than normal.

Edit:

No argument there ... thanks.


I think the biggest thing is the matchmaker. It's clearly working by both ECM and weight results.

#62 hammerreborn

    Member

  • PipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 3,063 posts
  • LocationAlexandria, VA

Posted 12 April 2013 - 09:10 AM

Is there anything else people would like me to examine? It just has to be something readily available from the data provided.

EDIT: I'm thinking about doing win rate of classes, and if I feel like really torturing myself, win rate of each variant.

Edited by hammerreborn, 12 April 2013 - 09:23 AM.


#63 hammerreborn

    Member

  • PipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 3,063 posts
  • LocationAlexandria, VA

Posted 12 April 2013 - 05:59 PM

So, since the devs post on matchmaking said:

Quote


WEIGHT CLASS MATCHING

We're adding the ability for us, the developers, to adjust our weight matching tolerances on the fly. This is giving us a controlled way to tighten up the weight imbalance between teams that players are experiencing.

Fun fact: About 74% of games kicked off are within what we consider tolerable weight matching limits (the difference between a heavy vs. assault Mech on the opposing team). Of the remaining 26% of matches, about 7% are what we would call "horrendously bad" (i.e., they carry the weight difference of one or two full assault Mechs between teams).


I decided to see how much my data correlates with their numbers.

And the result:

Posted Image

3% off if <35 tons is what they consider acceptable, or 6% if <30 tons. My over 100 tons matched exactly at 7%

#64 Kageru Ikazuchi

    Member

  • PipPipPipPipPipPipPipPip
  • The Determined
  • The Determined
  • 1,190 posts

Posted 12 April 2013 - 06:16 PM

I'm guessing they're using AS7 vs. DRG (40 tons difference) as their standard for "acceptable".

#65 Team Leader

    Member

  • PipPipPipPipPipPipPipPipPip
  • 3,222 posts
  • LocationUrbanmech and Machine Gun Advocate

Posted 12 April 2013 - 06:29 PM

Sorry that I posted and said I would try to help you out and then didnt... I had a busy week

#66 Odanan

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 8,210 posts
  • LocationBrazil

Posted 12 April 2013 - 06:32 PM

View Posthammerreborn, on 10 April 2013 - 08:43 PM, said:

Here are the results:

First of all thank you to all community contributers:
Rigatoni
Jay35
MuonNeutrino
Hauser
Kasiagora
RoBroCop
Krzysztof z Bagien
Jace Lancer
aniviron
Lance425
ElMarkoLoco
and
ClannerFodder

You are all the best
============================================================
So a couple people have asked how I collected the data and what it all means.

Below this post are links for each set of data (indicated by the letters in the top column A-0) aside from my own. Those links go to screenshots submitted by my awesome community contributers.

Each set of data is 10 rounds of assault, solo played, 8v8 matches without disconnects only.

Each contributers mech is ignored for the overall results (aside from their mech weight which is added to the overall tonnage.

The data is arranged as follows in the main chart:

Mech variant, mech total, %overall - should be self explanatory, each variant is seen total amount of times, and is % of the 2250 mechs counted.

ECM games, ECM win, ECM win % - If there was a game with a difference in ECM, it was an ECM game. If te team with more ECM won, it was an ECM win, and ECM win % is the ECM Win/ECM games

Heavy win, heavy difference, heavy average: If the team with more tonnage won, it was a heavy win, heavy difference is the absolute average difference over 10 rounds (including equal), and heavy average is the average weight of all the differences.


And now for the data
------------------------------------------------------------------------

Combined Results Raw Data!

Spoiler


Note: Heavier win rate % is not on the chart, it is 48.7% (roughly, there were 5-10 games where the weights were exactly even and are included in the overall number)

Note2: The heavier win is when the heavier team won, out of 10 matches (equal weight matches were counted as a loss).

Note3: Average difference for the weight is the average of the absolute value of each weight difference (so each game will have roughly a spider swing on either side)



Mech Diversity in Ranked form with Bonus Chart!

Spoiler


Class/Variant/Overall Breakdowns!

Spoiler


With Graphs!

Spoiler



Added 4/12: Weight distribution!!!

Spoiler

--------------------------------------------------------------------------------------------------
Impressions:

The three things that stand out the most to me is the huge numbers of founders and hero mechs that were fielded this weekend. Clearly people were out in force for these bonuses. We also see that the Heavy Metal, which many said was overpriced, absolutely DOMINATED the results with a whopping showing of 6.8% overall (remember, thats out of 2250 mechs counted!). If we were to assume these numbers are correct overall, and using the 100k number given for the be a hero weekend, that means there were roughly 6800 pinktarts flying through the skies, earning PGI about 240 thousand dollars. CRAZY!

The second thing that stood out for me is the Atlas D founders numbers. Holy hell. For the other founders mechs, all were roughly equivalent to their standard variant (Jenner D were both 24, the 4Gs 12/24, C1 35/38), and the Atlas went 35/108, making it nearly as played as the D-DC (and combined were actually played more)!!! I guess you could say that ECM is worth less than a 50% c-bill bonus, lol.

And lastly, as much as you see complaints on the forums about the matchmaker, the weights are roughly even (the maximum was 135 tonnage difference), and the win rate of the heavier team is roughly 50%, which seems to actually suggest it's doing something right.

I'd also like to make a shout out to Osakashogun, who was the lone Cicada 2B, how sad (for the variant).
-------------------------------------------------------------------------------------------------------------------


I hope you all enjoyed this as much as I enjoyed typing it, lol.

Raw data from each individual set of 10 to come in the following posts.

For those who are interested in my previous results for comparison, all the links to all of my studies can be found http://mwomercs.com/...ost__p__1672291

When compared to my March results (note that class based weight matching was present during this period), this is what appears!

Spoiler


------------------------------

For the poll data: please elaborate and feel free to discuss your choices, especially in regards to questions 1 and 2.

If anyone would like to have any of the raw/finalized data, I know someone has expressed that he wanted to upload the data on google docs when I get in touch with him. Or you can pm me directly and id be glad to share if you wanted to examine anything in particular on your own (like win rates for certain mechs. I just ask that you please give credit where credit is due. Typing 2250 mechs individually was pretty brutal.

Lastly, please keep all discussion civil.



Great info! (but your pool is confusing)

#67 hammerreborn

    Member

  • PipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 3,063 posts
  • LocationAlexandria, VA

Posted 12 April 2013 - 06:55 PM

View PostOdanan, on 12 April 2013 - 06:32 PM, said:



Great info! (but your pool is confusing)


What is confusing? Honest question. I tried to make everything as explained as possible.

#68 hammerreborn

    Member

  • PipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 3,063 posts
  • LocationAlexandria, VA

Posted 13 April 2013 - 08:31 AM

Added new poll question as to when I should do the next study.

#69 StandingInFire

    Member

  • PipPipPipPipPip
  • Bridesmaid
  • Bridesmaid
  • 152 posts

Posted 13 April 2013 - 09:01 AM

Like it my only problem, I want more graphs cause I am to lazy to read raw data.

#70 p4g3m4s7r

    Member

  • PipPipPipPipPip
  • Little Helper
  • 190 posts

Posted 13 April 2013 - 09:47 AM

View Posthammerreborn, on 11 April 2013 - 07:37 PM, said:


And because I'm awesome and have all the numbers in front of me.

Games with 2+ ECM: 36
Games where ECM won: 19

Win rate: 53%

Games with 3+ ECM: 3
Games where ECM won: 0

Win rate: 0%

So....there's that. More ECM actually lowered chance of winning.



You've really only shown that last weekend skewed the ECM pool so badly that we can't gain anything form this data (with regards to ECM). As you stated, anymore than 2 ECM per match probably indicated a premade last weekend. I'm confident this isn't the case normally. Also, only 3 matches with 3 ECM is too little data to really gain anything from analysis.

Also, one thing that is happening over time is that people are abandoning any light other than the 3L. The simple fact that the only light you aren't 3x less likely to see than 3L is a trial mech speaks to its brokenness. Also, since the trial mechs are likely new players, this means that they probably aren't going up against as many 3Ls, (granted your data neither shows nor disproves this).

After the 3L and heroes (and the founders jenner and, for some reason, the Jenner D) every other light mech and cicada has ECM. I think this is also going to skew your data in favor of balanced ECM, because matches with a team without any ECM will be substantially less common.

Also, it shouldn't matter whether or not a significant ECM balance is improbable. You should not accept imbalance just because a highly imbalanced scenario is improbable. It's still highly imbalanced, and still significantly less than fun.

I love this data though. I love statistics, and I think you've changed my mind regarding whether or not PGI actually does statistical analysis. However, I think they are coming to the same conclusions as you because they are only looking at a percentage win rate just for matches with more ecm, and not trying to pick out specific match types for analysis. You aren't really able to do that with your data because it's limited, so it's not like I'd expect you to come up with a different conclusion than them.

Edited by p4g3m4s7r, 13 April 2013 - 09:50 AM.


#71 hammerreborn

    Member

  • PipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 3,063 posts
  • LocationAlexandria, VA

Posted 13 April 2013 - 01:06 PM

View Postp4g3m4s7r, on 13 April 2013 - 09:47 AM, said:



You've really only shown that last weekend skewed the ECM pool so badly that we can't gain anything form this data (with regards to ECM). As you stated, anymore than 2 ECM per match probably indicated a premade last weekend. I'm confident this isn't the case normally. Also, only 3 matches with 3 ECM is too little data to really gain anything from analysis.


Uhhh....how so? If the argument is that more ECM = greater win it doesn't matter if there are a thousand matches with only a difference of one compared to 300 with 2 or more. Does ECM give a greater advantage, and the answer is barely 4% increased win rate, with all else not considered.

Quote

Also, one thing that is happening over time is that people are abandoning any light other than the 3L. The simple fact that the only light you aren't 3x less likely to see than 3L is a trial mech speaks to its brokenness. Also, since the trial mechs are likely new players, this means that they probably aren't going up against as many 3Ls, (granted your data neither shows nor disproves this).


The 3L has always been the most played light since its release. Mainly due to ECM + streaks which are still absolutely broken when used against lights. Cheese is used more than non cheese mechs, none of this is surprising.

Quote

After the 3L and heroes (and the founders jenner and, for some reason, the Jenner D) every other light mech and cicada has ECM. I think this is also going to skew your data in favor of balanced ECM, because matches with a team without any ECM will be substantially less common.


All the raw data is there, there were very few matches without ECM entirely. And your argument doesn't make sense because if lights are less frequent that means chances or seeing ecmless teams is less frequent?

Quote

Also, it shouldn't matter whether or not a significant ECM balance is improbable. You should not accept imbalance just because a highly imbalanced scenario is improbable. It's still highly imbalanced, and still significantly less than fun.


That's just like, your opinion man. I'm not even sure what you're trying to prove. "We'll, I'm only saying that a team of 8 ECM will crush a team with 0, and it totally doesn't matter that it only happens .00000000001% of the time but its still completely broken.

I'm going to pose the argument now that HMs are overpowered and not fun because there was 1 match possibly with 8 on a team and completely crushed the other team.

ECM is fine.

Quote

I love this data though. I love statistics, and I think you've changed my mind regarding whether or not PGI actually does statistical analysis. However, I think they are coming to the same conclusions as you because they are only looking at a percentage win rate just for matches with more ecm, and not trying to pick out specific match types for analysis. You aren't really able to do that with your data because it's limited, so it's not like I'd expect you to come up with a different conclusion than them.


Wtf am I reading? What else matters than the win rate? And what specific match types? Am I suppose to go through each map and list them so you can go "haha! On forest colony ECM had a 55% chance, totally overpowered!"

Edited by hammerreborn, 13 April 2013 - 01:38 PM.


#72 shotokan5

    Member

  • PipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 550 posts
  • Google+: Link
  • Locationvirginia

Posted 14 April 2013 - 05:34 AM

A nice step in the right direction. However does it take into account hud problems,crashes in game. Joysticks that are not fully operational. Going back to launch? Also, a number of other problems that are not addressed here. How can it be balanced if these problems still exist?

#73 Fate 6

    Member

  • PipPipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 2,466 posts

Posted 14 April 2013 - 05:52 AM

I was curious, but I'm not going to spend 10+ minutes matching the columns labeled A-Z with their data. Just so you know for the future.

#74 hammerreborn

    Member

  • PipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 3,063 posts
  • LocationAlexandria, VA

Posted 14 April 2013 - 08:40 AM

View Postshotokan5, on 14 April 2013 - 05:34 AM, said:

A nice step in the right direction. However does it take into account hud problems,crashes in game. Joysticks that are not fully operational. Going back to launch? Also, a number of other problems that are not addressed here. How can it be balanced if these problems still exist?


No crashes because no DCs are recorded, and as far as I'm aware there are no 0 damage dealers.

And as hud bugs are random they don't really mean anything in the overall balance. It's not like the entire enemy team gets it every game.

And how can what be balanced if there are problems? Mech choice and tonnage matching have nothing to do with any bugs or other "problems" with the game.

#75 p4g3m4s7r

    Member

  • PipPipPipPipPip
  • Little Helper
  • 190 posts

Posted 14 April 2013 - 04:57 PM

View Posthammerreborn, on 13 April 2013 - 01:06 PM, said:

Uhhh....how so? If the argument is that more ECM = greater win it doesn't matter if there are a thousand matches with only a difference of one compared to 300 with 2 or more. Does ECM give a greater advantage, and the answer is barely 4% increased win rate, with all else not considered.


My argument is that the advantage provided by ECM grows exponentially with the extent of the imbalance. Also, you only showed 3 matches with more than an ECM imbalance of 1. So you can't have any confidence in data regarding matches with an ECM deficit greater than 1. Also, I apologize, I should have said that my take on the data is that the 3050 weekend skewed the statistical representation of mechs in games too much to provide highly reliable data regarding the value of ECM. I think my first sentence sounded super douchy because I wasn't careful with my wording.

View Posthammerreborn, on 13 April 2013 - 01:06 PM, said:

The 3L has always been the most played light since its release. Mainly due to ECM + streaks which are still absolutely broken when used against lights. Cheese is used more than non cheese mechs, none of this is surprising.


But is the ratio of 3Ls to any other light variant 3:1 during any normal week/weekend? I guess my presumption was that people were probably playing lights less then because it was the 3050 weekend and the light pilots who stuck with their lights probably did so because they were more concerned with wins than cbills (but on second thought this may not be the most reasonable conclusion).

View Posthammerreborn, on 13 April 2013 - 01:06 PM, said:

All the raw data is there, there were very few matches without ECM entirely. And your argument doesn't make sense because if lights are less frequent that means chances or seeing ecmless teams is less frequent?


My whole point was that the 3050 weekend meant less frequent lights (except the death knell) which meant less ECM which meant a lower likelihood of substantially stacked amounts of ECM on either side or even very high ECM counts on either side. This means you'll have a less diverse of ECM match types to pull data from.

View Posthammerreborn, on 13 April 2013 - 01:06 PM, said:

That's just like, your opinion man. I'm not even sure what you're trying to prove. "We'll, I'm only saying that a team of 8 ECM will crush a team with 0, and it totally doesn't matter that it only happens .00000000001% of the time but its still completely broken.

I'm going to pose the argument now that HMs are overpowered and not fun because there was 1 match possibly with 8 on a team and completely crushed the other team.

ECM is fine.


I think there would be a problem with the Heavy Metal if 8 man heavy metal teams were able to consistently completely dominate any team they came up against, given that they'd be going up against another team of 8 assaults. I mean, this would kinda ruin 8 mans if people started abusing this hypothetical imbalance. And you know video games on the internet... people will abuse things that can be abused.

My point regarding improbable scenarios is that there is a component in the game where people are able to create whatever scenario seems highly improbable in a PUG match in an 8 man drop, or even sometimes in a sync drop. You could certainly say that other 8 mans should simply counter by doing the same thing, but this would reduce match diversity, which is counter to the intent of the game's design. So, I believe you are wrong, and that this is not just my opinion. I believe the devs think they are attempting to reduce any potential imbalance, no matter how unlikely the match-up is.

View Posthammerreborn, on 13 April 2013 - 01:06 PM, said:

Wtf am I reading? What else matters than the win rate? And what specific match types? Am I suppose to go through each map and list them so you can go "haha! On forest colony ECM had a 55% chance, totally overpowered!"


I'm sorry, I wasn't clear. This goes back to my point about significant amounts of ECM disparity between two teams. I believe data regarding matches with greater and greater amounts of ECM disparity between two teams would show greater and greater win percentages for teams with more ECM. There was a post a ways back that had a substantial number of ECM match types (4:1, 3:1, 2:1, 1:1, etc...) that argued that, statistically, the data supported an 80% confidence that ECM did in fact influence the outcome of a match and that this influence was in the favor the ECM team. The data and analysis were both very sound.

Anyway, I apologize again if it seemed like I was attacking your data. I really appreciate the work you did and like your presentation of the data. This kind of data is incredibly valuable to the community at large, and it bothers me that the devs make absolutely no effort to present it, as I've probably said many times before. It really leaves us dependent on people like you who put substantially more work than should be necessary to get us a snapshot of what the behavior of the game and community are in general.

#76 hammerreborn

    Member

  • PipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 3,063 posts
  • LocationAlexandria, VA

Posted 14 April 2013 - 09:50 PM

View Postp4g3m4s7r, on 14 April 2013 - 04:57 PM, said:

My argument is that the advantage provided by ECM grows exponentially with the extent of the imbalance. Also, you only showed 3 matches with more than an ECM imbalance of 1. So you can't have any confidence in data regarding matches with an ECM deficit greater than 1. Also, I apologize, I should have said that my take on the data is that the 3050 weekend skewed the statistical representation of mechs in games too much to provide highly reliable data regarding the value of ECM. I think my first sentence sounded super douchy because I wasn't careful with my wording.


If by 3 you mean 36 then yes....

There were only 3 games where there was a difference of 3 or more.

Quote

But is the ratio of 3Ls to any other light variant 3:1 during any normal week/weekend? I guess my presumption was that people were probably playing lights less then because it was the 3050 weekend and the light pilots who stuck with their lights probably did so because they were more concerned with wins than cbills (but on second thought this may not be the most reasonable conclusion).


3Ls have always been vastly overplayed. Last month the Raven 3L was at 4.6% played overall with the Jenner F falling behind at 1.9, the D at 1.3, and the 4x at 1.1. That one also had far less mechs accounted for (896) and was also prior to elo based matching (which stopped class v class matching), the srm hotfix (which lessened the value of streaks), and the laser state rewind (which boosted the value of the Jenner F).

Quote

My whole point was that the 3050 weekend meant less frequent lights (except the death knell) which meant less ECM which meant a lower likelihood of substantially stacked amounts of ECM on either side or even very high ECM counts on either side. This means you'll have a less diverse of ECM match types to pull data from.


There was just as little matches in March where there were substantial (not sure what you mean by substantial but they way you say it is "enough to prove my own perceptions") differences in ECM, where 3Ls were more likely to be present than they were this weekend. You just don't get games that often where one team ends up with 4 or more ECM, unless it's a sync drop, in which case the argument is moot because you have 8 people working together to stomp a PUG. I mean, you can even probably prove it statistically, with the numbers I've given, the chance of having 4 ecm mechs, have them all be on the same team, and have 0 on the other. It's just statistically improbable.

IIRC there were the same amount of 3+ games, 3, and none with 4 or more. My data wasn't written the same way as I did these so comparing the two now, especially this late, isn't worth my time, to be honest.


Quote

I think there would be a problem with the Heavy Metal if 8 man heavy metal teams were able to consistently completely dominate any team they came up against, given that they'd be going up against another team of 8 assaults. I mean, this would kinda ruin 8 mans if people started abusing this hypothetical imbalance. And you know video games on the internet... people will abuse things that can be abused.


8v8 doesn't have any weight matching at all. I mean, it's already being done. Have you done 8s? There are plenty of 8 D-DC, 8 3D, 8 3L, 8 cent 9-A teams running amok. Anything in large numbers is dangerous. Especially just due to the confusion it causes by having the same 8 mechs in the same paint schemes running at you. How do you know which to focus fire without cycling through every single one of them.

Quote

My point regarding improbable scenarios is that there is a component in the game where people are able to create whatever scenario seems highly improbable in a PUG match in an 8 man drop, or even sometimes in a sync drop. You could certainly say that other 8 mans should simply counter by doing the same thing, but this would reduce match diversity, which is counter to the intent of the game's design. So, I believe you are wrong, and that this is not just my opinion. I believe the devs think they are attempting to reduce any potential imbalance, no matter how unlikely the match-up is.


Not sure what you're trying to say here, but people will abuse whatever is currently the thing to abuse. Look at the fiasco both LRM bugs caused. Look at what happened prior to the SRM hotfix. Everyone was charging in with srms ablazing, splatterpults, etc.

But counters are always found in time. LRMs decrease, snipers rise, brawlers decrease, etc. it's always been an ebb and flow, and I really hope PGI never decides to balance a weapon based on "improbable" events. Flamers are fine because if 8 people run all flamer mechs than it can maybe cause a shutdown and kill people!


Quote

I'm sorry, I wasn't clear. This goes back to my point about significant amounts of ECM disparity between two teams. I believe data regarding matches with greater and greater amounts of ECM disparity between two teams would show greater and greater win percentages for teams with more ECM. There was a post a ways back that had a substantial number of ECM match types (4:1, 3:1, 2:1, 1:1, etc...) that argued that, statistically, the data supported an 80% confidence that ECM did in fact influence the outcome of a match and that this influence was in the favor the ECM team. The data and analysis were both very sound.

Anyway, I apologize again if it seemed like I was attacking your data. I really appreciate the work you did and like your presentation of the data. This kind of data is incredibly valuable to the community at large, and it bothers me that the devs make absolutely no effort to present it, as I've probably said many times before. It really leaves us dependent on people like you who put substantially more work than should be necessary to get us a snapshot of what the behavior of the game and community are in general.


The problem is as I stated, the bigger that gap gets, you don't have a random set anymore, you have a sync dropped premade, and their skill and communication is far, far, far more a benefit to their team than the ECM they are carrying. In my previous studies where I specifically tracked premades they had nearly a 90% win rate. The only thing that dictated a winner more than having a premade on your team was having a dc, which resulted in a loss almost every single time.

You just can't statistically form a team where you get 4 ECM on one side and 0 without something being really screwed up and manipulated from the outside. It just doesn't happen.

Edited by hammerreborn, 15 April 2013 - 09:10 AM.


#77 p4g3m4s7r

    Member

  • PipPipPipPipPip
  • Little Helper
  • 190 posts

Posted 15 April 2013 - 07:10 PM

View Posthammerreborn, on 14 April 2013 - 09:50 PM, said:


If by 3 you mean 36 then yes....

There were only 3 games where there was a difference of 3 or more.


Well, I guess I should stop complaining about ECM and move on to complaining about BAP (as in buffing it), TAG, and NARC. If they're supposed to be on par/ton then there's a serious problem.

I still think it would be better overall if they nerfed ECM, as opposed to buffing everything else (TAG could probably be buffed enough by proxy if they made LRMs better), but apparently their system is stable enough that it's not necessary.

#78 Damon

    Member

  • PipPip
  • 43 posts
  • LocationWyoming

Posted 15 April 2013 - 07:14 PM

Wow, nice data collection and analysis. Keep up the good work!

#79 Hammertrial

    Clone

  • PipPipPipPipPipPip
  • 267 posts

Posted 15 April 2013 - 07:16 PM

View Postp4g3m4s7r, on 15 April 2013 - 07:10 PM, said:

Well, I guess I should stop complaining about ECM and move on to complaining about BAP (as in buffing it), TAG, and NARC. If they're supposed to be on par/ton then there's a serious problem.

I still think it would be better overall if they nerfed ECM, as opposed to buffing everything else (TAG could probably be buffed enough by proxy if they made LRMs better), but apparently their system is stable enough that it's not necessary.


I fully support buffing BAP, but I think TAG is perfect where it's at now.

I learned today that I can do IF(AND()) functions to automate the ECM/weight winners, so that's awesome. That means the only manual thing I might have to do is input the mech data, then make the pictures, and combine. And the sorting afterwards, though there's probably some way to do that easier as well...

#80 p4g3m4s7r

    Member

  • PipPipPipPipPip
  • Little Helper
  • 190 posts

Posted 15 April 2013 - 07:30 PM

View PostHammertrial, on 15 April 2013 - 07:16 PM, said:


I fully support buffing BAP, but I think TAG is perfect where it's at now.

I learned today that I can do IF(AND()) functions to automate the ECM/weight winners, so that's awesome. That means the only manual thing I might have to do is input the mech data, then make the pictures, and combine. And the sorting afterwards, though there's probably some way to do that easier as well...


I agree that TAG is basically perfect, but the system it does such a good job supporting is terrible, which hurts TAG.

Excel is a wonderful thing. Especially when you get to the point where you literally need the extra memory provided by Office 2010 :D





14 user(s) are reading this topic

0 members, 14 guests, 0 anonymous users