349 replies to this topic

[Scurro]

I apologize if anyone received errors at random today on the leaderboard. I was performing some data migration. Everything should be complete now and no errors should be encountered.

Thanks!

[Scurro]

By request, a feature has been added to search results. You can now filter by season in group searches.

[Bombast]

Scurro, on 05 March 2018 - 03:29 PM, said:

By request, a feature has been added to search results. You can now filter by season in group searches.

I can't seem to search for Season 20.

[Scurro]

Bombast, on 05 March 2018 - 03:56 PM, said:

I can't seem to search for Season 20.

Thanks for the heads up. It is now fixed.

[Bombast]

Scurro, on 05 March 2018 - 04:07 PM, said:

Thanks for the heads up. It is now fixed.

How often to people tell you you're awesome? I bet it's not enough.

[UnofficialOperator]

So any eta on the overlay?

[Scurro]

UnofficialOperator, on 05 March 2018 - 08:20 PM, said:

So any eta on the overlay?

I am not developing that, but the overlay work is being done here: https://github.com/h...coreboardHelper

[UnofficialOperator]

Scurro, on 05 March 2018 - 08:50 PM, said:

I am not developing that, but the overlay work is being done here: https://github.com/h...coreboardHelper

Nice! thanks for all the work on leaderboards. Its great.

[Scurro]

I've added a new page of all player stats in MWO as well as I've started to include performance graphs for single player searches.

Enjoy.

https://leaderboard.isengrim.org/stats

[0Jiggs0]

Great work! Thanks!

[EnochsBook]

Yay! Graphs!

[Siegegun]

Wanted to say thanks for this tool. The work on it is appreciated.

Out of curiosity, is it possible in the future to add the ability to search and isolate by class? In other words, the ability to see just mediums or just lights and see the charts and results for just that class? Right now it is a global performance, which is great, but it would be cool to see how someone performs in just a single weight class. Lots of people are much better at certain classes than others.

If this is planned already, sorry I didn't read the entire thread. If not I think it might be a good feature to add.

[Scurro]

Siegegun, on 16 March 2018 - 04:39 AM, said:

Wanted to say thanks for this tool. The work on it is appreciated.

Out of curiosity, is it possible in the future to add the ability to search and isolate by class? In other words, the ability to see just mediums or just lights and see the charts and results for just that class? Right now it is a global performance, which is great, but it would be cool to see how someone performs in just a single weight class. Lots of people are much better at certain classes than others.

If this is planned already, sorry I didn't read the entire thread. If not I think it might be a good feature to add.

It's 100% possible as I have that information in the database, it would just require me to find a way to show it without cluttering up the website UI.

[Dr Cara Carcass]

UnofficialOperator, on 05 March 2018 - 08:20 PM, said:

So any eta on the overlay?

Taxxian from 12DG is developing that thing further to get higher OCR hit rates. For us it works quite well, perhaps write him in game, i guess he might like it if people find his work useful.
I will give him a heads up, that people asked for it here.

[Blue Pheonix]

Champion of Khorne Lord of Blood, on 06 September 2017 - 06:12 PM, said:

Still looking for a leaderboard that makes a distinction between solo and group queue.

AFFIRMATIVE!!!!!!!!

This time x1000

Matchmakers goal is to get a player as close to a W/L ratio of 1/1 as possible. It does this based on ranking/seeding players. Player ranks are based on tier level and other factors. These guys with over a 2.0 W/L ratio and substantial K/D ratios either have not played a significant number of games or are most definitely running on a team of at least one other person.

There is no way someone playing a substantial amount of games, over a long period of time, playing only 100% random PUG games with 100% random people can maintain over a 2.0 W/L ratio and that substantial amount of K/D ratios. Matchmaker is not that kind and the games are not that uniform and consistent. There will come a point that that good player will hit the MM virtual wall and will begin to be paired with players of lower skill until that player can not carry anymore and that players W/L ratio falls back closer to 1/1. This will be evident that this is happening when this player begins to notice that his teammates die off quickly most matches until W/L ratio gets closer to 1/1. MWO is a team game and just like with Lebron James, or Michael Jordan, while they are tremendously impactful, no player can solo a health team of 6 to 12 players on a consistent basis over a long period of time. It just is not happening. After many games played, trust me on this.

There is a huge difference in stats from someone on a team of at least one other person (who can coordinate, focus fire and kill over teamspeak) and random 100% PUG games with 100% random people. No way matchmaker is that kind and games are that consistent over many games and over a long period of time.

PGI and Jarls list, I know its been asked many times before but please provide a way to differentiate players who play on a team of at least one other person and those who play 100% PUG games with 100% random people (like myself).

Edited by Blue Pheonix, 16 March 2018 - 09:44 AM.

[Bombast]

Blue Pheonix, on 16 March 2018 - 09:33 AM, said:

Matchmakers goal is to get a player as close to a W/L ratio of 1/1 as possible.

That is how some MMs, for example WoT's MM, work. It is not how MWO's matchmaker works, though. It is entirely possible to maintain some fairly ridiculous stats in solo QP because 1) the MM makes no real attempt to curb streaks, and 2) because the metrics it does use (Namely, PSR) are stupid.

[Blue Pheonix]

Bombast, on 16 March 2018 - 09:40 AM, said:

That is how some MMs, for example WoT's MM, work. It is not how MWO's matchmaker works, though. It is entirely possible to maintain some fairly ridiculous stats in solo QP because 1) the MM makes no real attempt to curb streaks, and 2) because the metrics it does use (Namely, PSR) are stupid.

While I believe MWO Matchmaker is indeed flawed to a degree, after a substantial amount of games played I personally believe that MWOs matchmaker does try to get a player as close to a 1/1 W/L ratio as possible and that streaks are ended because of it.

I have noticed that once the Leaderboard updates after an hour of a strong W/L ratio streak, if played consistently, the next hour is often filled with games that lead in lopsided losses and the losses begin very early in the match with teamates dying quickly. If one plays long enough, they will start to see trends very early on the start of games and that suspicion/observation is usually validated by the statistics at the end of the game on the match score screen. A skilled player can usually tell the outcome of a game very early on in the game. This outcome and projected/estimated skill level of teamates is evidenced by multiple factors. From team spacing, team aim and actively hitting shots, team strategy, team movement etc.

I personally believe games are decided by the MM before games even start. This is by design and seeding of players. It evident to me based on how games play out from very early on in the game until the end of the game and from the streaks of losing and wining based on how one played the previous hour/last time leaderboard was updated.

I only have my personal experience to explain this.

The other request I have is for PGI to show "average match" score of each player playing that specific mech on the final screen at the end of the game. I have requested this before but it has not been implemented.

Edited by Blue Pheonix, 16 March 2018 - 09:58 AM.

[Grus]

Bombast, on 05 March 2018 - 03:56 PM, said:

I can't seem to search for Season 20.

Its up right now, just looked at my stats for that season.

[EnochsBook]

Blue Pheonix, on 16 March 2018 - 09:54 AM, said:

While I believe MWO Matchmaker is indeed flawed to a degree, after a substantial amount of games played I personally believe that MWOs matchmaker does try to get a player as close to a 1/1 W/L ratio as possible and that streaks are ended because of it.

I have noticed that once the Leaderboard updates after an hour of a strong W/L ratio streak, if played consistently, the next hour is often filled with games that lead in lopsided losses and the losses begin very early in the match with teamates dying quickly. If one plays long enough, they will start to see trends very early on the start of games and that suspicion/observation is usually validated by the statistics at the end of the game on the match score screen. A skilled player can usually tell the outcome of a game very early on in the game. This outcome and projected/estimated skill level of teamates is evidenced by multiple factors. From team spacing, team aim and actively hitting shots, team strategy, team movement etc.

I personally believe games are decided by the MM before games even start. This is by design and seeding of players. It evident to me based on how games play out from very early on in the game until the end of the game and from the streaks of losing and wining based on how one played the previous hour/last time leaderboard was updated.

I only have my personal experience to explain this.

While I don't question your personal experience, I really doubt that is how the MM works.

I've read the claim that the Matchmaker tries to push everyone to a 1/1 W/L R before, but as far as I know, the only thing PGI told us about the MM is that it's working off of player tiers.

It's possible there is more to it than that, but for the MM to make predictions as to who will win the battle beforehand it would need to have access to everyone on either team's stats (which it certainly could), everyone's loadouts (which it probably doesn't), and then use those two factors to determine with some degree of accuracy who is most likely to win.
Stats do play a huge factor in which team will win (eg the team with the highest average match score stands more chances of winning), but in my opinion a team's composition in terms of 'Mechs and loadouts is just as important. Now, among the relative top of players, loadouts don't really matter because you can expect everyone's 'Mech to be built very well, there will be few people with bad loadouts. In the lower tiers however, a lot more people don't know what they are doing, so average match score matters less and whose team has the most solid 'Mechs starts to skew things.

The MM probably can't "see" the loadouts on each team so I don't think it can accurately predict anything. Note that that would also require the people who designed the MM to have played a lot of MWO and to understand its mechanics and gameplay very well from a competitive point of view. I don't know that that's the case.

Edited by EnochsBook, 16 March 2018 - 10:57 AM.

[Scurro]

Unit leaderboards was added to the Jarl's List.

https://leaderboard.isengrim.org/units

Many units were imported from MRBC.

If you notice a typo in pilot name it was because they were submitted to MRBC incorrectly.

If you would like your unit added, removed, or modified, you can submit here.