283 replies to this topic

[Kroete]

Nightbird, on 14 July 2020 - 11:22 AM, said:

You're not going to get a normal distribution from Jay Z's PSR system. It's not based on stats.

You have a normal distribution on playerskill,
if this is not reflected in the matchmaking,
you will have problems in the long run.

Edited by Kroete, 14 July 2020 - 11:31 AM.

[Nightbird]

Kroete, on 14 July 2020 - 11:29 AM, said:

You have a normal distribution on playerskill,
if this is not reflected in the matchmaking,
you will have problems in the long run.

Jay Z's PSR is not based on skill, it is based on skill*times*games_played.

[UnofficialOperator]

Nightbird, on 14 July 2020 - 11:28 AM, said:

Get stomped 10 times in a row like many Tier 3 newbies are and they inflate the WLR of skilled players and after quitting early they never show in the chart for the low range.

It seems fitting that the community dies by its own hands.

[Tarl Cabot]

UnofficialOperator, on 14 July 2020 - 05:41 PM, said:

It seems fitting that the community dies by its own hands.

Yes and no. I believe players wanted a reset but most likely expected a reseeding of the player base. I would have at least preferred it have been at least within the 3 middle tiers. NO one asked for new players to start in Tier 3, that is on PGI but then the 2 making the calls do not play this game.

[Nightbird]

Tarl Cabot, on 14 July 2020 - 06:58 PM, said:

Yes and no. I believe players wanted a reset but most likely expected a reseeding of the player base. I would have at least preferred it have been at least within the 3 middle tiers. NO one asked for new players to start in Tier 3, that is on PGI but then the 2 making the calls do not play this game.

I didn't read anything in Jay Z's proposal about not seeding to T3. If anything, the proposal required putting everyone into T3 otherwise 'zero sum' wouldn't work.

It's pretty clear by now the sim is pretty accurate. Take a look at what Jay Z's proposal is doing next to the previous PSR system. Take a look at how a dumb WLR system surpasses it in terms of quality. It's the community's responsibility for putting forward a non-improvement, and as a result the community has to live with it for the rest of MWO.

Edited by Nightbird, 14 July 2020 - 08:27 PM.

[thievingmagpi]

Nightbird, on 14 July 2020 - 08:11 PM, said:

It's the community's responsibility for putting forward a non-improvement, and as a result the community has to live with it for the rest of MWO.

Oops! You left out the part where you did not submit an appropriate proposal.

So that's mostly on you.

Edited by thievingmagpi, 14 July 2020 - 08:44 PM.

[UnofficialOperator]

thievingmagpi, on 14 July 2020 - 08:44 PM, said:

Oops! You left out the part where you did not submit an appropriate proposal.

So that's mostly on you.

Lets put aside all the enmity. If we could redo everything and now that you have seen the current PSR, would you support Nightbird's work for the next update to PSR basing it on WLR only? Or how would you change the current PSR?

[MadcatX]

thievingmagpi, on 14 July 2020 - 08:44 PM, said:

Oops! You left out the part where you did not submit an appropriate proposal.

So that's mostly on you.

It's all good, PGI didn't choose an appropriate proposal.

PGI: Here are you're options: 1A, 1B, 2A, 2B "From what I've been reading, there's a slight majority of players who are favoring Core 1A or 1B above."
Community: *votes*
PGI: We're going with option 2C

If this is about scope... no... lets not pretend PGI cares about "scope" and use it as some sort of counter-argument. I remember when the game's primary "scope" was meant to be FW.

Edited by MadcatX, 14 July 2020 - 10:18 PM.

[Sjorpha]

thievingmagpi, on 14 July 2020 - 08:44 PM, said:

Oops! You left out the part where you did not submit an appropriate proposal.

So that's mostly on you.

Yes he did, in fact he's made several proposals tailored to different levels of access and complexity. He's put in more serious work than anyone else in the community. When PGI declined certain types of queries he deviced a way for the formula to work around that without having to do any of the things they said was out of scope.

On top of that he's made simulations to predict the performance of different solutions over time, not just for his own systems but for JayZs model as well. None of the other people creating suggestions actually bothered to simulate their own model like that, which is mind boggling when you think about it. And now we are seeing that the Nightbirds simulation checks out as predicted, which should give anyone still supporting the chosen model some serious pause.

But in the end, is quibbling about who did what really that important?

Isn't the important thing here to arrive at a working solution rather than personal bickering?

Doesn't the fact that the chosen model already manifests all the predicted problems in the simulation, and very likely will not be an improvement to match quality over the old system, bother you in the slightest?

For that matter, doesn't it bother the creators and propagators of this model that it is already starting to prove itself dysfunctional? Shouldn't they want to react to that development in some way?

What is your goal in this discussion? Is it to win some kind of personal argument or is it to actually fight for the possibility of a better matchmaker?

Edited by Sjorpha, 15 July 2020 - 01:08 AM.

[OneTeamPlayer]

Sjorpha, on 15 July 2020 - 01:08 AM, said:

What is your goal in this discussion? Is it to win some kind of personal argument or is it to actually fight for the possibility of a better matchmaker?

Quite a few people contributed to the propaganda surrounding the implementation of this current system, they're in full damage control mode at the moment.

Keep in mind that this current option was sold as a W/L solution which i was excited about.

Then I had an underwater W/L up to T1 and rounded out the full bar T1 at .97

Most of the losses were really bad losses too, not just stomp but shutout stomp or stomps where i managed the one kill on the entire team.

We were sold a false bill of goods, simple as that.

[John Bronco]

Seems like the community is getting exactly what they asked for. A pure W/L system was rejected by the masses in favor of a system that places more weight on MS, because of the emotional argument "I played well but lost so shouldn't be penalized".

The result of a popularity contest was not based on logic. What a surprise.

[Nightbird]

LOL, Paul closed the valves to try to make the results better. It should have some effect but we can't compare to season 47 results anymore.

[Jackal Noble]

Nightbird, on 15 July 2020 - 03:18 PM, said:

LOL, Paul closed the valves to try to make the results better. It should have some effect but we can't compare to season 47 results anymore.

I love waiting 10 minutes for MM to stack me against an EVIL squad. much fun. smooches.

If this in fact results in a more competent setting, I'm all for it.

[Nightbird]

Jackal Noble, on 15 July 2020 - 05:24 PM, said:

I love waiting 10 minutes for MM to stack me against an EVIL squad. much fun. smooches.

If this in fact results in a more competent setting, I'm all for it.

My simulation of Jay Z's system is with the valves closed rather than open, so we'll see.

Been waiting an extra 5 minutes for each 4 min match though, so I can see this hurting pop very quickly..

[OneTeamPlayer]

Nightbird, on 15 July 2020 - 05:31 PM, said:

My simulation of Jay Z's system is with the valves closed rather than open, so we'll see.

Been waiting an extra 5 minutes for each 4 min match though, so I can see this hurting pop very quickly..

Same, and this is in US primetime with both EU and US servers selected.

Heck i've dropped matches with Nightbird, then immediately entered the queue post game and still had a 5 minute wait, which is weird since moving from match to match should make catching the tail of 24 players out of all the players in T1 and T2 easy.

This valve closing really highlights how low the population has become and in reference to nightbird pointing out this may hurt population very quickly- 3 matches convinced me to hang it up for another few weeks.

The battle quality was slightly better, but the time isn't worth it at all. I'd point out that this 5 minute wait time seems to be someone trying to prove a point because somehow we managed both reasonably balanced and fast populated (1 minute or less) matches at noon US weekdays in the previous iteration but that would be obvious to most people who've been playing consistently.

This reset must have pushed out a hell of a lot of players.

[Anomalocaris]

OneTeamPlayer, on 15 July 2020 - 05:41 PM, said:

The battle quality was slightly better, but the time isn't worth it at all. I'd point out that this 5 minute wait time seems to be someone trying to prove a point because somehow we managed both reasonably balanced and fast populated (1 minute or less) matches at noon US weekdays in the previous iteration but that would be obvious to most people who've been playing consistently.

This wouldn't surprise me a bit. I rarely waited more than 3-4 min at 2am PST US pre-merge with only US server selected, so that kind of wait is.....unusual.

-edit-
LOL, just looked at Paul's post. 1 Tier separation at this point means there are probably about 100 Tier1 and Tier2 players available at any given time. And T2 could be playing against T3. So you might have 4 matches possible until the gates open. Of course this is the wrong approach to better matches, because you're throwing 90% of players into Tier 3 instead of having a better Tier separation.

Edited by Anomalocaris, 15 July 2020 - 05:55 PM.

[Nightbird]

Edited by Nightbird, 15 July 2020 - 07:52 PM.

[Nightbird]

Looking at the current distribution of player Tiers shared by Paul, I don't think the MS calculation is performing. People are moving more slowly than in reality. Something I'll keep in mind if we do this again in the future.

[Tarl Cabot]

Not really. It takes games to move up into Tiers (actual players) while your charts (percentage) are using the available Leadership Boards, regardless of number of games played. There are approx 1300 players in Tier 1-2. Thus could also be a number of players with low number of games played but high win rate, players who have not played enough games to move back up into Tier 1-2.

[Cluster Fox]

Hey Nightbird, have you revised your Stomp criterion from the cumulative Std Normal when you plugged in actual player data?

While I get the same %chances as you for the same WL cumulative delta (Which I set between 200 and 1800), my random games using the actual player data (season 38 to 46) yields much less stomps than 35% and usually a smaller WL cumulative delta per randomly assembled team than your original simulation with fake T1 players.

Keeping the stomp gate at 47.5% only gives me about 27% of matches being stomps.

If I set the gate at (difference between random% & win% <= 37.5%) I get about 35% stomps.

However we've experienced that the PSR reset to 2500 increased the Stomp% significantly. The value after PSR reset might be around the 75% mark and I think that's quite conservative. Sadly Paul didn't give us stomp data, that would have been great to have, but maybe too telling.

Edited by Cluster Fox, 16 July 2020 - 09:16 AM.