196 replies to this topic

[dario03]

Krivvan, on 26 March 2018 - 07:33 PM, said:

The funny thing is that, for many people, playing matches in large groups of good players actually lowers their average match score since you're competing with everyone else on the team for it and the matches end far too quickly and very little damage ends up being farmed. When players wish to pump up their match score stats, they generally have to do it in solo queue or in a small group. Another reason why although the Jarl list is decent for a general idea about a player, its not very precise.

Yeah, depends on a lot of things but if its a large group of all good players then avg ms will probably go down. Some exceptions though, like if you are being aggressive but wait for your assaults then the assault guys will probably still do a lot of damage. And since you will probably be winning most of the time they get the benefits of that too. If the tonnage is there my Mad Cat II-B usually still does a ton of damage in big groups but my smaller mechs vary a lot since sometimes you just don't have the time to get the damage once the fighting starts.

Edited by dario03, 27 March 2018 - 11:53 AM.

[Brain Cancer]

Quote

According to Season 19 data averages are:

WLR: 1.08
KDR: 1.00
MatchScore:

• Lights: 197
  
• Medium: 229
  
• Heavy: 234
  
• Assault: 247
  
• Global: 226

And people wonder why the event minimum was 100/150MS. If the median is 226, half the playerbase has trouble getting much higher than that.

[Zergling]

UnofficialOperator, on 27 March 2018 - 08:54 AM, said:

wow seems low for matchscore...

It may sound elitist, but only the top 10% or so of the player base actually have any real idea about what 'works' in the game.
With so many players simply not understanding how to play well, the averages are dragged down to such a low level.

Sadly, the PSR Tier system actually disguises the true skill difference between the top players and everyone else.

Brain Cancer, on 27 March 2018 - 11:43 AM, said:

[/b]
[/list]And people wonder why the event minimum was 100/150MS. If the median is 226, half the playerbase has trouble getting much higher than that.

Amusingly, the average MS used to be even lower before things like the skill tree, with overall average being almost exactly 200.

Edited by Zergling, 27 March 2018 - 12:42 PM.

[Sjorpha]

Zergling, on 27 March 2018 - 12:42 PM, said:

It may sound elitist, but only the top 10% or so of the player base actually have any real idea about what 'works' in the game.
With so many players simply not understanding how to play well, the averages are dragged down to such a low level.

Sadly, the PSR Tier system actually disguises the true skill difference between the top players and everyone else.

While I agree that there is a huge gap between the good players and the rest in this game, it is not true that the average statistics are "dragged down" by this.

The averages are pretty much hardcapped by the limitations of how much damage there is to do, how many mechs there are to kill and how many matches there is to win etc. For W/L it's easy since the average stat across all players is necessarily exactly 1.00, but it's actually true for all the other stats as well.

If everyone played as well as the top 1% it wouldn't automatically increase the average stats, some of them (like damage and MS) might even go down due to more efficient shooting. Or maybe they go up slightly if better torso twisting/spreading outweighs the better accuracy I don't know, but in either case we're talking pretty small shifts.

The high stats of good players are possible ONLY because others are bad in comparison and play in the same matches for the good players to outperform.

For example the average stats in a WC final is not much different than the average stats of a tier 5 pure newbies match, the former is infinitely more high level play of course but that doesn't increase the amount of damage needed to win and so the averages remain more or less the same.

Edited by Sjorpha, 27 March 2018 - 01:46 PM.

[Solahma]

Sjorpha, on 27 March 2018 - 01:41 PM, said:

...but it's actually true for all the other stats as well.

Not true. The opponents you go up against can vary in HP, yet the value for killing any opponent is the same. Perhaps you only target assault mechs as a priority. You will deal a lot more damage for every kill that you get. Compared to only targeting lights and mediums that require a lot less damage to destroy. Yet the value for each kill is the same according to your K/D stat.

[Grus]

Brain Cancer, on 27 March 2018 - 11:43 AM, said:

[/b]
[/list]And people wonder why the event minimum was 100/150MS. If the median is 226, half the playerbase has trouble getting much higher than that.

Odin's beard that explains so much....

if the median is 226 then the min should be 225. Hoping the other side of the coin would get better... I would hope.

[Brain Cancer]

Grus, on 27 March 2018 - 04:17 PM, said:

Odin's beard that explains so much....

if the median is 226 then the min should be 225. Hoping the other side of the coin would get better... I would hope.

225 is actually an improvement from before, where MS tended to drift around 200. Again, much more than 150 and significant numbers of players would actually have trouble making the event happen.

That's kind of disturbing. The kind of games during the event where I didn't make 150, much less 100 were underhive-tier or trying to play that Roughneck hero without any real materials to rebuild it with, since I normally only play Clan chassis.

[Grus]

Brain Cancer, on 27 March 2018 - 04:26 PM, said:

225 is actually an improvement from before, where MS tended to drift around 200. Again, much more than 150 and significant numbers of players would actually have trouble making the event happen.

That's kind of disturbing. The kind of games during the event where I didn't make 150, much less 100 were underhive-tier or trying to play that Roughneck hero without any real materials to rebuild it with, since I normally only play Clan chassis.

I mean ok if you're in a unskilled much for the first time i don't expect you to pull a 12 kill right out of the box... but damn...even if you hit everyone on the other team (12 assists) wont you brake 200?

[Brain Cancer]

Grus, on 27 March 2018 - 04:30 PM, said:

I mean ok if you're in a unskilled much for the first time i don't expect you to pull a 12 kill right out of the box... but damn...even if you hit everyone on the other team (12 assists) wont you brake 200?

These are players who in some case, you'd consider doing better than usual hitting more than one target while still moving.

The Underhive is a very special place.

[Popcat]

Ok So a couple of quick points:

1)Jjarl if you bothering to read this. Thanks, I like being able to review my statistics across multiple seasons and have not been able to do so through MWO. Probably just haven't found the right spot yet.
2) is there any way to show a players clan mech vs IS mech ratio?
3) I always used my K:D ratio as my personal helping / growth measurement figure if I'm killing more and dying less it's probably my greatest influence in the out come. So I ask would this not be a better indicator of skill? why or why not?
4) Looking at my MWO stats does this mean they think I played 24855 days in my rifleman!!!! Well it is my "Baby" RIFLEMAN RFL-3N 135 48 84 0.57 21 113 0.19 20,605 73,196
24855 days 03:14:07

[justcallme A S H]

UnofficialOperator, on 27 March 2018 - 08:54 AM, said:

wow seems low for matchscore...

Believe it or not... It's actually high/gone up.

Before skill maze the average was around 30pts lower as an average. My guess as to why:

- Survival tree
- Huge mobilty nerfs from engine desync
- Many, many good players have stopped playing
- Large number of weapon nerfs and additional of inaccurate weapons like MRMs and RACs.

[UnofficialOperator]

Popcat, on 27 March 2018 - 08:28 PM, said:

2) is there any way to show a players clan mech vs IS mech ratio?

You can use the addon called Madstats for your browser

[Sjorpha]

Solahma, on 27 March 2018 - 02:41 PM, said:

Not true. The opponents you go up against can vary in HP, yet the value for killing any opponent is the same. Perhaps you only target assault mechs as a priority. You will deal a lot more damage for every kill that you get. Compared to only targeting lights and mediums that require a lot less damage to destroy. Yet the value for each kill is the same according to your K/D stat.

This is strawmanning my position, I'm not talking about the averages of individual pilots.

I'm talking about the global average damage to kill a mech for example, that average doesn't change very much depending on who targets which weight class in what order. It obviously affects your own average damage, which isn't what I'm talking about here.

Global damage average is basically capped by the total durability times average number of kills per match, the population is large enough that your individual choice to focus assaults won't make a noticable difference to these averages.

You can imagine model scenarios where everyone gets more accurate and the average damage goes down due to increased efficiency, or everyone twists better and the average goes up and so forth. Or your example where everyone suddenly decides to always go for a certain tonnage priority in targeting which would indeed increase the average damage to the extent it increased the percentage of lighter mechs amongst the survivors on the winning team (though there is already a big to small targeting bias in the game I think), it obviously makes no difference on the losing side in cases where the whole team dies.

In reality though, the averages are mostly defined by average durability x kills, as examplified by the much higher average damage scores on the clan side in Faction Play because there are more hitpoints to chew through.

Skill is a factor in the sense that accurate shots lowers the averages, though skilled tanking increases them, not sure if they cancel out or not. I'm inclined to beleive that there is both a higher skill ceiling and a faster progression rate for accuracy than there is for tanking, which would net a reduction in average damage (again per match, not individual) as you go up the skill brackets, but that's speculation.

You could test it by sampling enough low and high end matches and compare the average damage scores, my personal guess is that the average total damage dealt per match is higher in tier 5 than in tier 1 most of the time.

The bottom line is that there is no reason to believe that a more skilled player base would net higher average stats, it's just as likely (if not more likely) to decrease them a little.

Edited by Sjorpha, 28 March 2018 - 01:55 AM.

[Zergling]

Sjorpha, on 27 March 2018 - 01:41 PM, said:

While I agree that there is a huge gap between the good players and the rest in this game, it is not true that the average statistics are "dragged down" by this.

The averages are pretty much hardcapped by the limitations of how much damage there is to do, how many mechs there are to kill and how many matches there is to win etc. For W/L it's easy since the average stat across all players is necessarily exactly 1.00, but it's actually true for all the other stats as well.

If everyone played as well as the top 1% it wouldn't automatically increase the average stats, some of them (like damage and MS) might even go down due to more efficient shooting. Or maybe they go up slightly if better torso twisting/spreading outweighs the better accuracy I don't know, but in either case we're talking pretty small shifts.

The high stats of good players are possible ONLY because others are bad in comparison and play in the same matches for the good players to outperform.

For example the average stats in a WC final is not much different than the average stats of a tier 5 pure newbies match, the former is infinitely more high level play of course but that doesn't increase the amount of damage needed to win and so the averages remain more or less the same.

Fair point.

There is probably just a huge difference in skill level between the players that have some idea of what they are doing, and everyone else.

The PSR Tier system definitely disguises the extent of that skill difference though; it effectively deflates the stats of Tier 1 players, while inflating the stats of everyone lower Tier.

Brain Cancer, on 27 March 2018 - 04:26 PM, said:

That's kind of disturbing. The kind of games during the event where I didn't make 150, much less 100 were underhive-tier or trying to play that Roughneck hero without any real materials to rebuild it with, since I normally only play Clan chassis.

I hit 150 in almost all my battles for the event; I can only think of a 2 or 3 battles where I derped up hard or my team got utterly steamrolled and I failed to reach that score.

[Grus]

Brain Cancer, on 27 March 2018 - 04:42 PM, said:

These are players who in some case, you'd consider doing better than usual hitting more than one target while still moving.

The Underhive is a very special place.

Is that like the Darkweb or something? lol

[Brain Cancer]

Grus, on 28 March 2018 - 11:22 AM, said:

Is that like the Darkweb or something? lol

It's the depths of MWO derp. Places where builds make no sense, tactics do not exist, and even movement and firing together is often somehow taboo. It's where LB-Xs are considered good because they deal additional critical damage and players seem to play at random, apparently blissfully unaware of opponents that aren't red squares directly in front of them.

Like Camelot, it is a very silly place.

[Sjorpha]

Zergling, on 28 March 2018 - 09:40 AM, said:

Fair point.

There is probably just a huge difference in skill level between the players that have some idea of what they are doing, and everyone else.

The PSR Tier system definitely disguises the extent of that skill difference though; it effectively deflates the stats of Tier 1 players, while inflating the stats of everyone lower Tier.

Yeah, for sure. If we actually had a sensible ELO style system the rating would be public with enough granularity to see the difference all the way from bottom to the top, and we'd use the rating itself to see who's good or bad and not have to look at the stats. Players like Proton et al would have ratings like Magnus Carlsen and Grishuk has in Chess and the matchmaker could actually know the vast skill difference between for example me and them.

Edited by Sjorpha, 28 March 2018 - 04:12 PM.