349 replies to this topic

[Nightbird]

You can always tell though, by claims of how OP lights are and UP assaults are in this game lol!

[Eisenhorne]

Oh Jarl's can definitely be gamed, I played like an absolute scrub with LRM-only assaults and got to 99.5%.

I think the point is though that while being a high% doesn't indicate skill, it indicates that you understand the game well enough to game the score. A player who is unskilled won't understand the meta enough to know how to break it. I'm obviously not better than 99.5% of the people in the game (at least I don't think I am) but I gamed the score and got a high rank. Woo? Kinda pointless.

Conversely, a low score shows you either don't understand the game at all, or you do understand it very well and are actively making decisions that hurt your team. Like, if you build a 12 ERPPC direstar and try to one shot enemy mechs, you obviously did that on purpose and know what you're doing, and you'll have a very low score because of it. If you play poorly because you don't understand the game, you'll also have a very poor score.

I cannot speak for everyone, but I have personally never seen someone who has a rank below 60% that I'd consider "good" at this game, at least not good enough to speak on what is over or under powered.

[Nightbird]

Eisenhorne, on 13 August 2018 - 12:14 PM, said:

Oh Jarl's can definitely be gamed, I played like an absolute scrub with LRM-only assaults and got to 99.5%.

We keep asking PGI for separate solo and GQ stats. If we had solo quick play stats, W/L ratio will determine how good you are. You can be a damage monster, expert light back stabber, charismatic commander etc... anything you're doing to help your team win will be reflected in you WLR.

[Kubernetes]

MummyPig, on 12 August 2018 - 10:54 AM, said:

Ranking means virtually nothing tbh. Misses out where players take on specific roles that have little impact on k/d ratio etc. such as the need for ams/narc pilots to combat or work with the current messed up meta. Taking down 7.5k missiles but only doing 200 - 300 damage and no kills shows up as a poor result in this kind of analysis, yet gives a match score comfortably in the 500= zone and as a team player, is enabling ranked pilots to do their part of the team effort.

There are lies, there are outrageous lies, and there are statistics.

Look, another dude who reflexively dismisses Jarls despite not knowing the first thing about it.

Smh

Edited by Kubernetes, 13 August 2018 - 12:27 PM.

[Scurro]

As an update, there is a known issue where searching by some unit tags will result in errors. It is currently being worked on. Searching by unit name does not have any problems.

EDIT: Fixed

Edited by Scurro, 13 August 2018 - 06:34 PM.

[Jman5]

xX PUG Xx, on 13 August 2018 - 11:22 AM, said:

You are debating with someone I would and have gladly had at my side in many battles: we both know his limitations and skills and we both know I value him much higher than many "Top Percentile" players. He may not get the massive damage scores but he is a reliable ally and good pilot to have on your team for the very reason he has stated. We are not all armour farmers and cockpit takers

Oh and my names PUG, go ahead and tell me I'm not worth debating with

I don't really care if someone is good or bad at a game. I happen to be pretty good at MWO, but I know there are 1000 other games I've played over the years that I'm absolutely awful at. Like you, I would much rather play with someone with a likeable personality than some a-hole 1%.

I also don't really care if people want to delude themselves about how none of the stats are capturing their true value to the team. While I think this can happen, there are usually clues in the data that indicate it. (such as higher WLR than you'd expect, and/or a large amount of games as a light indicating he may go hard for low-scoring objectives)

What I don't like is when someone comes in here and calls the work Scurro has done meaningless. While, I think there are many variables we just can't account for, I do think jarl's list gives us a rough approximations. Personally I lump all the top 1,500 player into the same category.

Quote

EDIT

Just to clarify, I use the List to track my own stats but I also know I can farm the crap out of match score to significantly boost those very same stats. Bringing 3xAMS on a KFX and standing next to couple of big slow assaults netted a rather impressive 500+ match score.... with only 2x ERMLL fitted to my 'Mech.

The problem with any stat based system is that it can be "gamed" if all you are interested in is your position relative to others and not improving you're actual game play in relation to a normal match. In my opinion anyway

I would be curious to see just how well a triple AMS kitfox could average in a season. ~4.5 tons is an awful lot of tonnage for a 30 tonner. Certainly you could have monster games when the missiles are flying, but would it be enough to counteract all those low scoring matches where there just aren't tons of missiles in the air?

My feeling is if triple AMS led to consistent high scores, it would be a bunch of randos taking top 20 every season instead of people playing more traditional, damage-dealer roles.

[S O L A I S]

Eisenhorne, on 13 August 2018 - 12:14 PM, said:

Oh Jarl's can definitely be gamed, I played like an absolute scrub with LRM-only assaults and got to 99.5%.

I think the point is though that while being a high% doesn't indicate skill, it indicates that you understand the game well enough to game the score. A player who is unskilled won't understand the meta enough to know how to break it. I'm obviously not better than 99.5% of the people in the game (at least I don't think I am) but I gamed the score and got a high rank. Woo? Kinda pointless.

Conversely, a low score shows you either don't understand the game at all, or you do understand it very well and are actively making decisions that hurt your team. Like, if you build a 12 ERPPC direstar and try to one shot enemy mechs, you obviously did that on purpose and know what you're doing, and you'll have a very low score because of it. If you play poorly because you don't understand the game, you'll also have a very poor score.

I cannot speak for everyone, but I have personally never seen someone who has a rank below 60% that I'd consider "good" at this game, at least not good enough to speak on what is over or under powered.

Um, you gamed the match score. That is not in the control of the Jarl's list but even so remains one of the best indicators of player skill. The Jarl's list is there to show us the seasons in one spot and hey what do you know it shows that your leap 99.5 is not that far off than the progression you had over the previous two seasons.

While your damage may have been inflated by lurming, you still quite likely outperformed most. You see knowing where to be, and how to be effective with what you are using is skill. Not all players can jump into an LRM assault and produce consistent enough results to game themselves into the 99th percentile.

Edited by S O L A I S, 14 August 2018 - 12:03 AM.

[General Solo]

I Like Jarls list as it gives me a quick indication who's advice is better.

So many times some dude is telling me how good they is

500 every game bro

I take them to the group q and its sub 100, sub 200, 50 in an assualt mech

Than I check Jarls list and say

No wonder , that makes sense
Edit: Data comes from PGI

Edited by OZHomerOZ, 14 August 2018 - 04:19 AM.

[Nightbird]

Could someone post a link to the spreadsheets again? I'm interested in re-crunching the numbers

[Scurro]

Nightbird, on 20 August 2018 - 03:22 AM, said:

Could someone post a link to the spreadsheets again? I'm interested in re-crunching the numbers

https://drive.google...ZGc?usp=sharing

[K O Z A K]

It's funny how despite not playing qp/gq for a couple of months now my ranking is somehow improving, lol. I guess the game is really bleeding high end players

[Grus]

can someone get a hold of the vikings and let them know their sight is down?

[Scurro]

Grus, on 20 August 2018 - 05:00 PM, said:

can someone get a hold of the vikings and let them know their sight is down?

The leaderboard? It is up for me. What error are you getting?

[Grus]

Scurro, on 20 August 2018 - 05:07 PM, said:

The leaderboard? It is up for me. What error are you getting?

it was the one where it showed 3 different connections. on of my connection to the net the other to the cloud and the other to the jarls list. had a big red X on the list.

Edited by Grus, 20 August 2018 - 09:03 PM.

[Blue Pheonix]

Jarls list is a great option compared to the standard MWO leaderboards. However, as some have said, it still does not differentiate between solo play with 100% random people and play and when you drop with at least one other person consecutively and communicate tactics, kills and the like through Teamspeak (or another means). This is a big deal.

There is a tremendous difference between completely random PUG play (which I do/did) and teaming up with at least one other person. Ive done both but 99.99% of my games are with 100% random people. I can tell you that their is a huge difference between the two. The players who team up will almost always have significantly better stats when pairing with another player. Significantly better. Especially with W/L ratios.

The whole point of MM is to get two teams as close to equal as possible. Like it or not, players are weighted based on their W/L ratios, damage, K/D ratio and the like. The team will consist of players who perform consistently high in those categories and then seeded off of those players.

If a player plays 100% PUG games with 100% random people and plays in a high enough tier with a high number of game played, eventually all players will hit the virtual wall. They will be able to climb to a point until they are eventually paired with a team that consists of players who dont play very well. This will be evident in the way ones team plays - teamates dying quickly early in the match and upon observation, making poor choices (worse then average). At that point, it doesn't matter how good a player is, no player can solo a team of 6 - 12 healthy mechs. So, W/L ratio takes a hit. Again, the MM goal is to get it as close to 1/1 as possible. In addition to W/L taking a hit, with that lopsided of losses and a mismatched team, K/D and other stats take a hit. Again, its a vitual wall.

This is the reason why I stepped away from the game for a few months now. Too stressful, especially when you are competitive. A good player will still have over a 1/1 W/L ratio with a significant number of games played as they carry to some degree but again, there comes a point the virtual MM wall is hit and no matter how hard a player tries to carry, its mathematically impossible to win due to heat, DPS, time and opposing skill factors. Unless the other team just decides to all be terrible shots all of a sudden or power off for no reason

But yes, players can game Jarls list by playing with at least one other player. Doing so and playing with 100% random people is not differentiated. When I see guys with 3.0 - 6.0 W/L ratios with a decent number of games played, I know instinctively they are playing most, if not all (or at least a significant portion) of their games with another player who they drop with. Based on how the current MM works and the Virtual wall I talked about, there is no way in H - E - double hockey sticks that a player can maintain that level of W/L ratios on completely random PUG play with 100% completely random people over a significant number of games. With that exorbitant of a W/L ratio, of course K/D, damage and all other stats will be higher too. Just my $.02.

What is needed, and it has been repeated many times before, is a way to differentiate stats from completely random PUG play where a player is dropped with 100% completely random people versus the player who drops with at least one other person intentionally, even if it is only occasionally. MWO and Pirahhna could do this as it would be simple enough to track, they just are choosing not too. For instance, If player "Hodog and ketchup" drops with player "hamburger and pickles" and they are on the same team 3 (or more) matches in a row, this is not random coincidence, they are deliberately dropping together. Those games they dropped together would be tracked under "group play" and the games either of the two dropped with 100% random people would be tracked under "solo play". Again, not hard to track, just not a priority to implement the tracking of it.

Anyway, Ive been out for a bit and just cruising the forum to see if any major changes (nerfs, buffs, additions, deletions) have been made in the last 3 or 4 months before I jump back in.

Cheers

Edited by Blue Pheonix, 24 August 2018 - 01:33 PM.

[MischiefSC]

Scurro, on 20 August 2018 - 05:07 PM, said:

The leaderboard? It is up for me. What error are you getting?

Wait, so your sight is down? Someone hide your runes? Doesn't surprise me. Actually the idea of Jarls List being a product of predestination, created by a blind and implacable force of fate against which there is no remedy is a pretty reasonable way to look at QP matchmaking for most people.

So since you're clearly in with 'The Girls from Urðarbrunnr', are they more like the Völuspá or like 'Ah! My Goddess!'

That makes a big, big difference on which is hotter, Skuld or Urd and which one having a poster of is creepy.

[Jackal Noble]

I had to look up UROBRUNNRRRRR. have a like.

[DrxAbstract]

OZHomerOZ, on 14 August 2018 - 12:17 AM, said:

I Like Jarls list as it gives me a quick indication who's advice is better.

So many times some dude is telling me how good they is

500 every game bro

I take them to the group q and its sub 100, sub 200, 50 in an assualt mech

Than I check Jarls list and say

No wonder , that makes sense
Edit: Data comes from PGI

Homer why u no luff me?!

[O L L O]

Scurro said:

Thanks yet again for another month, Scurro - the jarl's great tölva skald!

Edited by Ollollo, 02 October 2018 - 05:02 AM.

[Scurro]

Just a small update:

I've since adjusted the database to now match PGI's leaderboard for seasons.

This was leading to confusion when at some point PGI changed season 1 to season 0.

Everything will now match.

For a future update I plan to add some timestamps near seasons.