Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[dougz]

Starting with tonight's update to the leaderboard (if all goes well), only the past 30 days of games are going to be counted towards your rank.  Games older than that will have no impact whatsoever on your score.

This means that if you get better, your old games won't be dragging down your rank forever.  It also means that you can't maintain a high rank based solely on past performance, playing only an occasional game to keep your account active.  I think this is a good thing — I'd rather a high rank mean an active, fresh player — but you may disagree.  If there's something really wrong with it I can always restore the old method.  This is my third attempt at trying to account for player's skills varying over time; if nothing else it is at least the simplest to explain.

Yes, Karumah is still on top (and his sock puppet Kretin is now dead last, by far), but I expect everyone who cares knows about him so really, who cares.

[painted_cow]

Well, personally I dont like this. Players who play a smaller amount of games per day were kind of "punished" with a higher deviation skill number with todays (or old) method. Now people, who play many games per day get even more advantage or am I wrong?

Or isnt there a chance to do an alternative all-time-Leaderboard then (also without the like 10 days inactivity kicks of players). Would be funny to see the whole number of players on isotropic as well and remember some oldtimers.

[theory]

Seems like this will cause some pretty swingy ranks.  A level 40 player who doesn't play for a month is suddenly going to come back as a level 0, and start doing real damage to everyone's ratings.

In general it will probably depress everyone's level except for those that play pretty much every day.  I worry that this might overly reward activity in place of skill.

[painted_cow]

In general it will probably depress everyone's level except for those that play pretty much every day.  I worry that this might overly reward activity in place of skill.

Who is seeing paralles to Magic the Gathering here? :-) Planeswalkerpoints anyone

[theory]

What if you set it to the last X games a player played, instead of the last 30 days' worth of games?

[rod-]

Last X or a number of days >30 would be good.  I'd enjoy having my first 400 or so games fall off of my ranking, but i've probably played 80 games in the last 30 days.

[DG]

This means that if you get better, your old games won't be dragging down your rank forever.

Although a player's ranking is determined from all the results, surely the mathematics of the rating system already guarantees that the contribution of the latest results is far higher than the contribution of the older results? Even if a commutative series of deltas are added to the rating with each game, wouldn't the size of the deltas add more weight to recent results?

[DsnowMan]

I don't like it. Echoing DG, I thought the rating system already had some time-decay.

If you want to something like this, I also suggest using X # of games, not days.

[heatthespurs]

I believe you need to consider that there could be quite a number of player (like me) who play only 1-2 games per day on average. If the rating is based only on games in last 30 days, the rating would be constantly based on around 40-50 games. I wonder would this give a constantly high variance that would not be reduced over time? If yes, it may not give a fair reflection on the player actual skill, even over a long time?

I second that the rating could instead be based on last X games, or games on last Y days with Y>30

[kn1tt3r]

Maybe it is a bit late since the system might be shut down in a year or whatever, but I'd personally like to see the results on a regular basis ("seasons"), like a league systems just measuring the skill level in the span of maybe quarter of a year.

If you'd additionally could change the view to some sort of All-time ranking, which could only show, say, the top 500 without disappearing names in terms of inactivity, it would be perfect.

However, in the end I don't care very much, and I'm in principle fine with whatever change done with good intentions.

[Elyv]

I don't like it; I tend to play in spurts, so I'll play very little(or not at all) for a month or two, and then play a ton for a couple of weeks. I find I can still hold my own against people my level fairly well once I start again, and I'm certainly still better than a level 0 even after an extended period of inactivity.

[DStu]

I would think for the average player that should not change much. The old games diminish in TS already in 2 ways (afaik). The system does not weight the current rank with the number if games, so if you are say 25+-10 with 100 games or with 1000 makes no difference for your future. Second, when you, for historical reasons, are below your 'real' skill, say 25 instead of 35, the system will adjust your rank for a much high amount when you win again the same player, because the skill difference between him and you is more in your favour.

So there are the border cases, first the ones that exploited theirselves on top. I'm still not sure if they can maintain there rank by just playing 1 game per week, because that would increase the variance of the skill in the long run. But of course the new method gets rid of them faster.  But probably we will see some more creative skills to exploit the leaderboard in the future.
Second, the case mentioned by theory, having someone good not playing (much), dropping to 25+-25 by this. Hmm.

[Kirian]

What if you set it to the last X games a player played, instead of the last 30 days' worth of games?

Have to second theory's idea here.  Someone who plays less than, say, 100 games in a month is going to end up with a rank no higher than, say, level 30.  On average, likely less.  Consider that people who gamed the system (e.g., Paralyzed) took 100+ games to do it.

I think the last 300 or 500 games is a better indicator than the last month.

[DStu]

OK, I must correct myself and say the opposite: Have fallen from 34 to 23 (and even increased from 21 when you calculated yesterday ranks with the new method).
AND I must say that definitely my rank had increased (in the old system) during the last month.

[octopus]

Here are some relatively easy ways to get a high rating:

1) play only low-level players
2) play only 3p/4p games, which implies mostly low-level players; I did this, now I'm number 5 (yay?)
3) veto cards you're no good with
4) veto boards that look scary to you
5) veto players you don't like the look of

Here are some ways to fix the problems:

1) Only count games where players are all +/- 5 levels
2) Create separate 2p, 3p, 4p leaderboards
3) Don't count veto mode games
4) Only count games with a fully blind kingdom
5) Only count blind automatch games

Yes, there's some overlap there, but I think those changes would make it much harder to game the system.

[Lionel]

It doesn't change anything. I like playing dominion before this leaderboard change, i like playing dominion after this leaderboard change. That is the most important no?

[theory]

The problem with my idea (basing it on the last X games) is that you could very easily end up winning a game and going down in rank, which doesn't really make much sense.

I thought about it last night, and it seems that there's not much options beyond doing a "Top Players in Last 30 Days" leaderboard and a separate "Top Players All-Time" leaderboard.  Outside of TrueSkill's built in time decay, if you want to further encourage activity and highlight the recent stars, there's not much options other than what doug has done here. 

At the same time, the top ten of today's leaderboard has only 2 of yesterday's all-time top ten (plus Karumah, of course).  I'm currently level 27, and whoever I play with is going to have their ranking seriously distorted.  This ends up defeating the original purpose of TrueSkill, which was to promote fair and automatic matchmaking. 

[theory]

Here are some relatively easy ways to get a high rating:

1) play only low-level players
2) play only 3p/4p games, which implies mostly low-level players; I did this, now I'm number 5 (yay?)
3) veto cards you're no good with
4) veto boards that look scary to you
5) veto players you don't like the look of

For what it's worth, Captain_Frisk did this with his conjecture account and made it pretty high.  I think the problem is just that the system's initial assumption of skill at 25.0 +/- 25 is too generous.

[Geronimoo]

Is this change a convoluted method to erase the Paralyzeds of the community from the leaderboard?  (the guy who played nothing but the KC-Masquerade pin)

I sure don't like this change and think the leaderboard works fine as it is.

[rspeer]

Oh god, I really don't like this change (and I'm the one who told dougz about trueskill in the first place).

This is throwing out tons of information, it's rewarding being obsessed with the game, and the worst part is the idea that you lose your entire rating if you don't play for 30 days. And this certainly didn't help the Karumah situation.


Incidentally, the 25 +/- 25 assumption isn't generous, it's arbitrary and it defines the way the rest of the numbers come out. If newbies got 10 +/- 10, then the system would be the same except elite players would be level 18. I think what's actually going on is that the number representing the amount of variance inherent to Dominion needs to be turned down. Unskilled players play with high variance, and therefore the average game has high variance. But skilled players can play with low variance, especially against unskilled players, and we notice most how the leaderboard affects skilled players.

[Caprica]

I really dislike/hate the change! The system worked fine.

It would be cool though, to have the all time leaderboard and a leaderboard based on the last 30 days. But before that's reality, I want the old leaderbaord back!

[DG]

I suspect that while this might spur people in the top 20 places in the rankings to play more games it will probably discourage the other 6000 players. The ranking system will mainly be measuring the time invested in playing an on-line game over a short period, which isn't a measure that many people want to know. For those people who have invested a very large amount of time over that time period it will then measure skill playing Dominion, just for that time period.

[Mean Mr Mustard]

As an upshot, I suppose those of us who lost long-standing ranks will now be mirrored less by players who do not know us.  My standings have declined quite a bit since Hinterlands so I guess I don't mind, but I can see that I would be kind of annoyed if I had still been at the top and lost my position.

After playing 10,000 games (including a heavy stretch at BSW) I have grown a bit bored with Dominion, but as a guy with a lot of financial obligations it is hard to give up an absolutely free form of entertainment.  My play style has relaxed and I have become much more experimental and my results have been reflecting that.  Last week Geronimoo chastised me for trying to circumvent Mountebank, and I had to laugh because he was dead-on right.

I suppose that ultimately I assumed that Isotropic has an unseen self-destruct mechanism installed, and that the leaderboard would go with it, so I might as well not worry about my rank.  The shadow game that is supposedly forthcoming will provide a clean slate, and if I choose to buy it I expect to see a lot of the same people. 

Now, I can understand Dougz's attempts to keep the leaderboard sanitized of outright manipulators and I also get the frustrations of established players who may be losing rank from infrequent play.  I also am having a few doubts that a commercial Dominion online product release is eminent.  But what do I know?  It is all guesswork and speculation at this point.  I am not in the inner-circle, but I can tell you as someone who organized online tournaments I did consider the possibility of the downfall of Isotropic when planning them, and I see the DSC tourney as a good omen.  From DonaldX, the last I remember hearing was that Isotropic would be up through Hinterlands release.

Pragmatically, I guess if this leaderboard change is permanent we must play in order to defend our rank. 

[Reyk]

In German there is a saying "to throw the baby out with the bath water" (in the sence of "to overshoot the mark"). Why not just manually delete "Paralyzed", "Karumah" and the likes?

Anyway thx again for hosting isotropic, dougz!

[Sopenas]

Hello,
 
  In my opinion 30 days period used for counting rank is too small. I think it should be about 3 month time - 90 days. Then it would be ok. Now you go on vacation and after month you are at zero level.  But your skill definitely is not at zero level after months holiday 

  This ranking system would be too unstable.

Bye

[Fabian]

Let me add my voice to the list of players who greatly dislike this change. The reasons have already been outlined above by other posters, but please consider them.

[painted_cow]

Delete Karumah and I am ranked #1 again (nice change one could think)

But I dislike the system eventhough! The argument, that games like 6 months ago should not influence todays skill is, as someone already mentioned, in the "old" system of Trueskill. Take an example. Lets say I lost 1 game 6 months ago, or maybe 10 games in a row. This drags down my Trueskill number for a short time. But in addition my future games after that losses will be rewarded with higher gainings. So in the long-run games that long ago wont matter anyway. Hope this argument is understandable :-)

[Fabian]

Can we at least please have the old leaderboard back too This seriously is making me sad (admitedly probably more than it should though).

[Geefour]

I wonder which date will be chosen as the basis of the seeding for the upcoming tournament.  I was looking at playing one of the top-ranked players in the first round.

[Qvist]

The problem with my idea (basing it on the last X games) is that you could very easily end up winning a game and going down in rank, which doesn't really make much sense.

I thought about it last night, and it seems that there's not much options beyond doing a "Top Players in Last 30 Days" leaderboard and a separate "Top Players All-Time" leaderboard.  Outside of TrueSkill's built in time decay, if you want to further encourage activity and highlight the recent stars, there's not much options other than what doug has done here. 

At the same time, the top ten of today's leaderboard has only 2 of yesterday's all-time top ten (plus Karumah, of course).  I'm currently level 27, and whoever I play with is going to have their ranking seriously distorted.  This ends up defeating the original purpose of TrueSkill, which was to promote fair and automatic matchmaking.

I don't see the problem with this. If we count the last X games, every game one game drops out of your rating. So let's say, that game that dropped out of my ranking was a win against a Lvl 40 player, and now I win against a Lvl 30 player, of course the rank goes a little bit down, but that shouldn't change so much. That's probably not more than -0.05 considering 500 games (just an estimation).
And it could happen so now too. If you play today one match and win against a Lvl 30 player, but 31 days ago you played exactly one match too and won against a Lvl 40 player, your rating drops too.

Most of the sport rankings work this way.
http://en.wikipedia.org/wiki/FIFA_World_Rankings
http://en.wikipedia.org/wiki/ATP_Entry_Ranking

The problem is you are now forced to play massively. If not you get high variance. That seems not fair.
I would leave it as it was because ranking systems like ELO or TrueSkill are designed to be balanced.
If it's because some people found out how to tweak it, then not the system failed. Then you have to change which games to take into account like already suggested.

[Görling]

+1 for reverting back to the old scoring

I do, however, think that imposing stricter rules about what games contribute to your ranking could be fruitful.

IMO The definition of being truly skilled at the game is your ability to beat any player on any board. Therefore i suggest that the only games that should contribute to your ranking are tournament games and automatches with unbiased boards and no veto. Also turning down a suggested automatch should make you lose a small amount of ranking points.

A compromise could be to have random games contribute more to your ranking than veto-games or something along those lines.

I see no reason to temper with peoples current ranking. If Karumah is such a bother to you why dont you just manually reset him or whatever?

[Stealth Tomato]

Also not a fan. I liked jockeying for position on the leaderboard--now it just doesn't seem right, especially with the ridiculous variance coming in.

It was also nice being able to see from the leaderboard approximately how long someone had been playing (career ranked games).

[Kirian]

The problem with my idea (basing it on the last X games) is that you could very easily end up winning a game and going down in rank, which doesn't really make much sense.

True now anyway, since the rankings are only calculated once a day.  If I had a really good day 30 days ago, and I only play once today for a quick win against someone low-level, my ranking will drop.  THis makes it necessary to play lots of games... which isn't really the point of Isotropic, I think.

It's also (of course) now possible for rankings to drop without playing any games at all.

The analogy I thought of is this:  consider what would happen if FIDE did this for chess players.  The outcry would be instant and impressive.

Now... we played without a leaderboard for many months.  It could easily go away, and we could go back to playing whomever.  But I think the intent of the leaderboard was to be able to find people within a reasonable range of your own skill.  This completely blasts that out of the water.  I mean... theory at rank 27?  MMM at 27?  That's nuts.

[Marcus316]

Hmmm ... seems like there are a lot of theories about how the leader board should work. I'll throw in my thoughts ...

First, concerning gaming the system: I'm only just starting to read through details about TS, but it seems to me that perhaps the goal definition for the algorithm should be more complex than who wins, perhaps taking into account some sort of priority of goals (in the simplest case, winning by pile depletion vs winning by resign). A game played to the end is far more reliable a judge of skill over a win by resign. Is TS flexible enough to make the distinction ... and does the game log data provide enough detail to recalculate based on the two different end-states? I think this is far more likely to help with Resign-bot cheating.

Second, concerning stale skill vs current hot players: there's no easy answer. On some of the Go servers I play on, the last 6 months worth of games are taken into account, and there's a large diminish in relevance as a game falls back in months. Overly active players still have issues with stale ranking, but it usually clears up if they sow down for a week or two. The other item that helps is to provide an option for playing non-ranked games with ranked opponents, allowing two additional benefits: 1) players don't have to stop playing to ease the staleness of their rank; 2) high-ranked players can freely choose to play non-ranked games against other high-ranked payers in order to experiment with new or crazy strategies, or just to relax with a friend.

Just my few thoughts.

[nsnyder]

I agree with everyone that 30 days is awful short.  Dominion is a game with very high variance.  If we're going to sunset old games it should be something more like last 1000 games or 3 months whichever is longer.

I'm not totally thrilled about the compression of the rankings.  My relative ranking hasn't changed (near the bottom of the top 100) but the number has gone from 37 to 31.  Seems like there's less info there.

[Lekkit]

I think that the leaderboard change should not matter to most of the people. Either you take the game and the leaderboard seriously, and if you do you should probably be able to play a few extra games. Or you play for fun, and then it doesn't affect you.

[painted_cow]

@ Lekkit: This thought may only be true for players, who a) start right now and will play a huge amount of games or b) play a huge amount of games anyway. For established players, who played their biggest amount of games already and reached their goals its a pity not to be at the top anymore.

By the way, was here even one player who liked the changes?

[def]

So, why is there an option to only play other players within a certain level range? To make sure that you can chose to play players who probably have a similar skill level like you.
Isn't that a main reason for the leaderboard - not to see who is better, but who is similar good, which leads to pleasant games.
The new system contradicts this, since I won't know what the real level of my opponent is, as long as he isn't playing 100 games per month.

[biopower]

First, concerning gaming the system: I'm only just starting to read through details about TS, but it seems to me that perhaps the goal definition for the algorithm should be more complex than who wins, perhaps taking into account some sort of priority of goals (in the simplest case, winning by pile depletion vs winning by resign). A game played to the end is far more reliable a judge of skill over a win by resign. Is TS flexible enough to make the distinction ... and does the game log data provide enough detail to recalculate based on the two different end-states? I think this is far more likely to help with Resign-bot cheating.

There are so many games where people resign because they don't want to play the game in the end, or the opponent already got more than half of the VP on the board, or each turn takes too long because of pawn/hamlet. Maybe restricting the penalty to games where one player resigns before Turn 5 would be much better than just a flat penalty on all resign games.

[rspeer]

There are so many games where people resign because they don't want to play the game in the end, or the opponent already got more than half of the VP on the board, or each turn takes too long because of pawn/hamlet. Maybe restricting the penalty to games where one player resigns before Turn 5 would be much better than just a flat penalty on all resign games.

If it counts less to resign early, that would just encourage people to say "5/2 gg" and resign instead of playing the game.

[olneyce]

This change 'helps' me, I guess.  #17, wooo!  But it actually just makes cracking the top 20 feel a little cheap, since it's all in one jump, as opposed to the slow slog up the charts I had been conducting.

I don't think it's a problem to include the old games.  If people really want to be free from them, they can just start a new account.  I have one I started mostly just to experiment and mess around and it only took about 100 games to get it right up into the 30s.  But I enjoy playing with the whole corpus of work, including a bunch of stupid early losses, associated with my real account.

[Mergus]

I agree with many of the opinions posted in this thread so far. For me personally, it was a nice challenge to try and climb the old leaderboard and I like checking it for that reason. With the new system being able to advance will likely become more of a time than a skill challenge to players, which I think is not an improvement.

I also think that the leaderboard is mainly meant to give an idea about the skill of your opponent, in case you never played or heard of him. The new system will not do such a good job of this, unless you and your partner are both really active players. For example, if you look at today's leaderboard, shark_bait is Level 13 and in my opinion he's one of the best players, if you judge it by his success in the two Isodom tournaments. He just doesn't play a lot and with only 22 of his games counting, TrueSkill can't rank him higher.

I know that there are probably many ideas on how to improve the leaderboard and I think it's great that dougz has tried something new and you can always go back (which I would also encourage for the recent update).

An idea I had for a while was that the leaderboard could show your skill among the players that are actually listed. I think everyone agrees that it makes sense not to list players after they have been inactive for a while. Some of these players will come back and some of them never do. So what if the leaderboard would count all of your games (like in the old leaderboard) but only against players that were active in the past X days. I don't know how much longer it would take to update the leaderboard in such a way, but if it's easy to implement, I would really like to see the results. This would be a way to show your skill as it compares to the active community. It's just a question how you would define "active" in that case. If you use X=30 days for this method it probably won't change the old leaderboard too much.

Anyways, this is a good discussion, I hope the outcome will be as well

[rspeer]

Okay, this is more ridiculous than I thought.

Playing some casual Thanksgiving day games with my friends -- who may play less than I do but are by no means terrible at Dominion -- I find they're all level 0 and 1. The kinds of players who may not be well-represented here on the forum have been basically wiped off of the leaderboard.

Comment from one of them: "Well, I guess this teaches me about the impermanence of all good things."
Another: "You can fix this, right?"

[Lekkit]

I actually like the changes. Now it's a leaderboard where you can see who the best active players are. Before it was more like a hall of fame. Sure, 30 days seems like too little to base rankings on, someone mentioned 90 days, and I think I would prefer that.

Most of the "good" players read these boards anyway, and knows that Shark_Bait (just an example) is a good player. A fast-changing leaderboard is not something I see as negative. It encourages people to actually play the game more. Is that bad?

[Jorbles]

I agree with many of the opinions posted in this thread so far. For me personally, it was a nice challenge to try and climb the old leaderboard and I like checking it for that reason. With the new system being able to advance will likely become more of a time than a skill challenge to players, which I think is not an improvement.

I also think that the leaderboard is mainly meant to give an idea about the skill of your opponent, in case you never played or heard of him. The new system will not do such a good job of this, unless you and your partner are both really active players. For example, if you look at today's leaderboard, shark_bait is Level 13 and in my opinion he's one of the best players, if you judge it by his success in the two Isodom tournaments. He just doesn't play a lot and with only 22 of his games counting, TrueSkill can't rank him higher.

I think Mergus brings up a really good point here. I care a lot more about the overall skill level of the person I'm playing against than I do about how they've played in the last 30 days. Especially if they've only played a few games in the last 30 days.

[chwhite]

I'll forty-fifth the sentiment that this is a bad change and the old leaderboard was better in basically every way.

[The Adventurer]

I completely agree with all that was said on this topic (everything but that bad idea that started the topic, that is).


It is a bad idea to rely solely on this method, though it might be a fun idea to have as a supplement to the original leaderboard (by which all rankings should remain based on). Don't replace the old with this. Why do we have to pick one and only one? You could have all-time records (main), last 30-days, last 500 games, last... etc as complements of info for fun.

[boloni]

Why do we have to pick one and only one? You could have all-time records (main), last 30-days, last 500 games, last... etc as complements of info for fun.

I would also like different leaderboards. Maybe dougz can add a player-identifier to the game logs so that councilroom or other sites can uniquely identify a player and create alternative leaderboards.

[shark_bait]

I can't add that much to what's been said, but with the old leaderboard, I was enjoying snailing my way up the leaderboard with my limited time.  I don't really see how a ranking system that doesn't accurately portray the skill of a person can be a good ranking system. 

I'll echo what other's have suggested regarding a recent activity leaderboard and an all time leaderboard.  That way we can see what highly active players are the best and continue to see low activity veterans represented by an accurate level.  For now though, I'll just have to be one of the most fearsome Level 13's around 

[toaster]

I don't have much to add, except to agree that this change makes the rankings much less enjoyable and useful.  Here's hoping the change will be reconsidered.

[Dominionaer]

Can the TrueSkill algorithm get enhanced with a factor for number of opponents? Or that the second - third - fourth ... game against same opponent changes less?

[Davio]

How about using a Ladder system which resets every 30 (or 90) days to accommodate the more active players? Every player starts with 0 and just tries to work his way up.

Besides the ladder, the hall of fame persistent rankings can be maintained.

On the main page it can show something like:
Davio 14/33 or something with the first number being the ladder rank and the latter being the all-time rank.

[Lekkit]

On the main page it can show something like:
Davio 14/33 or something with the first number being the ladder rank and the latter being the all-time rank.

I like this idea.

[guided]

I wouldn't be that bothered except I haven't been playing much the last month and this torched my rating just in time for tournament seeding. A second all-time leaderboard would be nice. But yeah, the FAQ has always said "Okay, I promise not to take it too seriously." My approach to leaderboard 'sploiters like Paralyzed has ever been to roll my eyes and pay no attention to them.

[Donald X.]

Can the TrueSkill algorithm get enhanced with a factor for number of opponents? Or that the second - third - fourth ... game against same opponent changes less?

The simplest thing to do here would be to only rank games played against random opponents. If you pick your opponent, that game doesn't count. I don't know if this would bother people or what.

TrueSkill can only rank you within a community of players. If you just play one person then it's only ranking you vs. them. And other than the "look at me" guy, there are no doubt people who just/mostly go on isotropic to play friends who then get inaccurately ranked.

The most accurate thing would be to only show a list of you and the people you've played. Then you don't get a single leaderboard though. You could instead only rank people who have played against N people who have played against N people, as a way to guess whether you're really part of the community or not. You could still beat that approach by making tons of dummy accounts and having them all play each other, and maybe this is a community where it's hard to get your first N games because people want to know your ranking in order to play you; dunno there.

[Guy Srinivasan]

If you want a leaderboard, then this change is absolutely fine, but showing people their "ranking" when they're not in the top tier of players is silly - the leaderboard is optimized to show who's at the very tippy top. If you want a measure of skill, then using TrueSkill levels is a bit silly, since they're optimized for a leaderboard-like system. Otherwise subtracting uncertainty from mean skill as a "conservative measure of skill" doesn't make much sense to me. If I had my druthers then it'd show the TrueSkill mean and some sense of the uncertainty like "Rating: 30 (28-32)" vs "Rating: 30 (27-33)".

[Reyk]

The simplest thing to do here would be to only rank games played against random opponents.

The sad thing is that you have to pick your opponent after the new leaderboard was released. Because the auotmatch +/- level option no longer works in a meaningful way.

[painted_cow]

The sad thing is that you have to pick your opponent after the new leaderboard was released. Because the auotmatch +/- level option no longer works in a meaningful way.

Furthermore why should I play say a level 13 shark_bait or level 27 theory??? It would be completely dumb for me to do it. I can only lose points in a long run...Not really encouraging. So, I am still waiting for the good old leaderboard, please set it back, dougz :-)

[fellowmartian]

There seems to be a biggish groundswell of opinion to the effect that this is a bad idea. I said in a recent post elsewhere that you have to ask what a leaderboard is for. I think most people think of it as some combination of:

a) a challenge, a bit of fun to see how high you can go
b) a way to see how you are progressing in skill
c) a way to find similarly-skilled opponents.

The new system seems a significantly worse way to achieve all three of these. The only people who are going to be able to use it in this way are people who play very regularly. People who've been working their way up but have to have a break will see themselves drop away again; that's discouraging. People who've dropped away but are really good players won't show as such (if I want a level 27 and I get theory, (c) above is right out).

Add one more to the chorus of 'please change it back', or at least provide the two alternatives...

Or, if this is permanent, then... rrenaud, councilroom is separate, right? And uses all the same data? In which case would a separate leaderboard based around the old system be plausible on councilroom? Or something like this? We could put our councilroom levels in our statuses if needs be. I'm no techie so forgive me if this isn't a plausible idea.

[Razzishi]

I fail to see the point of this change.  Trueskill by itself already does whatever this change would possibly be intended to do if properly configured.  I think that the way of only adding to a player's uncertainty on days that they play games is a really strange thing to do, and that every player's base uncertainty should go up by a small amount each ranking period regardless of whether they play or not while playing games at a consistent skill level will naturally lower uncertainty.  That way people will slowly drift down the leaderboard when they're inactive, and Trueskill will work its magic moving them quickly to the rank that they should be when they return.

edit: You could then add a significant increase to every player's uncertainty when a new set comes out; I think that the changing nature of the game might have been a major part of rationale behind this change.

[popsofctown]

Can the TrueSkill algorithm get enhanced with a factor for number of opponents? Or that the second - third - fourth ... game against same opponent changes less?

The simplest thing to do here would be to only rank games played against random opponents. If you pick your opponent, that game doesn't count. I don't know if this would bother people or what.

TrueSkill can only rank you within a community of players. If you just play one person then it's only ranking you vs. them. And other than the "look at me" guy, there are no doubt people who just/mostly go on isotropic to play friends who then get inaccurately ranked.

The most accurate thing would be to only show a list of you and the people you've played. Then you don't get a single leaderboard though. You could instead only rank people who have played against N people who have played against N people, as a way to guess whether you're really part of the community or not. You could still beat that approach by making tons of dummy accounts and having them all play each other, and maybe this is a community where it's hard to get your first N games because people want to know your ranking in order to play you; dunno there.

You should know, though.  It's not.  Come play with us Donald

[rspeer]

CouncilRoom does have its own implementation of TrueSkill. I'm seeing if I can get a sensible approximation to the previous leaderboard out of it.

There is some missing information, unfortunately. We don't know whether the players were logged in or not, and we don't have unique IDs for their accounts -- so if a player changes their name, it would have to start their rating over.

Also, it's coming out in a different range somehow. Nobody ever ends up over level 30. Which might have been how the old old leaderboard worked. Does anyone remember what the change was that dougz made before this? I think it had to do with how uncertainty was added to ratings, such as adding the uncertainty factor once per day instead of once per game or something.

[theory]

Also, it's coming out in a different range somehow. Nobody ever ends up over level 30. Which might have been how the old old leaderboard worked. Does anyone remember what the change was that dougz made before this? I think it had to do with how uncertainty was added to ratings, such as adding the uncertainty factor once per day instead of once per game or something.

Yeah.  He used to increase it a little bit per game, but no one could gain enough points to overcome that.  So now it's a little bit per any day that you play at least one game. 

[guided]

I always wondered, what if it's per game but smaller? Or what if it's per day period but smaller?

The "per day that you play a game" has always subtly discouraged me from logging on if I don't have time to play a bunch of games.

[nsnyder]

The uncertainty tweak never really made sense to me.  The reason the "variation" part was high before is just that Dominion is a high variation game.

[nsnyder]

What I find frustrating about the change is that it used to be that I could look at how my rating changed to get an idea of how well I played the day before, but now it doesn't work that way because it's as much a reflection on how well I played 30 days ago as it is on how I played yesterday.  So it's actually less responsive to my play rather than more responsive.

[rspeer]

I always wondered, what if it's per game but smaller? Or what if it's per day period but smaller?

The "per day that you play a game" has always subtly discouraged me from logging on if I don't have time to play a bunch of games.

Adding uncertainty per day makes a lot more sense to me.

I'm experimenting with this right now: at the start of each day (or for now, each 15,000-game set, because I'm too lazy to try to parse dates out of the councilroom data), everyone's variance moves 1% back toward its default value.

If you play frequently and constantly, this is equivalent to the previous leaderboard. If you play frequently and then don't play for a week, your uncertainty increases (and therefore your level decreases) by about 1.7 points. While you're not playing, your level effectively decays with a half-life of 69 days.

I'll see how it comes out. Although I would be mostly happy with just having the on-Isotropic leaderboard back the way it was.

[popsofctown]

Wait, is NO ONE over thirty?

Is it possible that my 3 rank drop doesn't mean I'm sucking lately?

[Rjax36]

I feel sorry for my opponents in this new system

I'm lifetime lvl 36 (I've topped off at #16 on the leaderboard) and I play tons of paper dominion even when I'm inactive online, so the 5-20 games I play each month should be pretty miserable for my opponent's ratings...

I STRONGLY prefer the old system

[Axxle]

I find that currently I'm ranked about the same and haven't seen many problems with it.  But I can definitely forsee myself trailing off in activity like before and would be very discouraged if I had to crawl back up from rank 0 again.  I'd be very bored with playing people much below my rank in order to do so.  So count me in for reverting the change.

[Masticore]

So i think this is an interesting idea, but this way isn't giving it a chance. The time period is too small, it needs to be 3 months at absolute least, hopefully 6 months. For now i say change it back and then revise a plan to make it better

[Lekkit]

I was thinking about this earlier today, and I think that keeping levels and the leaderboard separate is the way to go. Levels could reflect overall stats, and the leaderboard could reflect recent achievments.

Generally I feel like people don't like the changes because their level doesn't reflect their skill. Keeping level on another board than the leaderboard would solve this problem, right?

[Dominionaer]

Generally I feel like people don't like the changes ...

I like my jump up. And can lie to me, that those ahead are not out of reach.

[Kirian]

The sad thing is that you have to pick your opponent after the new leaderboard was released. Because the auotmatch +/- level option no longer works in a meaningful way.

Furthermore why should I play say a level 13 shark_bait or level 27 theory??? It would be completely dumb for me to do it. I can only lose points in a long run...Not really encouraging. So, I am still waiting for the good old leaderboard, please set it back, dougz :-)

Exactly!  Think of this from a game theory perspective:

No reasonable person of high skill and high knowledge will play against shark_bait while he's level 13, if they're interested in a maintaining a position on the leaderboard--that is, if the board has any interest to them as a little bit of competition.  So a few different types of people will play him:

People with no interest in using the board for competition (who wouldn't anyway)
People who don't know his old ranking
People with very low rankings

The first and third will be essentially unaffected.  Two groups are affected:  those who won't play him but would have been fine playing him at level 40+, and those who don't know that old ranking.  The former lose out on an interesting opponent; the latter group have their scores disproportionately deflated.  Meanwhile, because higher-ranked people won't play him, his rating won't rise as fast as it ought to, perpetuating the problem.

I think fellowmartian has it right above.  It serves none of the likely purposes of a leaderboard, and most definitely kills what I think is the main purpose--to be able to play against those nearest in skill to you.  There's even a menu option to force auto-match to do that for you!  But that menu option is now... much less useful.

I basically have two options:  ignore the leaderboard (all right, I can do that).  Or--and this is worse--not play for a week or so and strategically deflate my rating in an attempt to inflate it by playing games against people who ought to be much lower than I am.  I'm not nearly as well-known as shark_bait or theory (except, perhaps, among members of this board).  It's not my style, I won't do it... which means basically ignoring the leaderboard for mow, which saddens me a bit.

Doug?  You there man?  Tons of top players begging for this change to be reverted, not for the sake of their ratings but for the sake of the system.

[rspeer]

Here's an experimental leaderboard with the "uncertainty decay" thing I was talking about, where the uncertainty range gradually increases toward 25 at a rate of 1% per day. There are probably glitches.

http://councilroom.com:8888/leaderboard

(This link will probably go dead in a few days. Hopefully because the code gets merged into the main site.)

Interestingly, the top player on it is samojedi, who if I recall correctly was the top player on the very first leaderboard. I had no idea he was still playing.

Also interestingly, the level range is still smaller than isotropic's (not to say that that's inherently bad). I wonder what other unpublished tweaks Doug has made to the rating code.

[rrenaud]

samojedi was crazy good.  I wonder who it was. 

[DStu]

Also interestingly, the level range is still smaller than isotropic's (not to say that that's inherently bad). I wonder what other unpublished tweaks Doug has made to the rating code.

I thought the isotropic Trueskill-code is on Github. Are there parameters missing?

[rspeer]

Apparently. Because if you run the code as-is, you get everyone having a mean skill between 15 and 35, putting all the really good players in a narrow range of levels from about 25 to 29. And it clearly doesn't contain other changes that have been made to the isotropic leaderboard, such as the change that adds uncertainty at most once per day.

It doesn't make sense to speak of samojedi in the past tense: he played 74 games yesterday. But he's not on the official leaderboard, which means he plays logged out.

I think this highlights a problem with running a leaderboard using the CouncilRoom data: there are players who play logged-out, either because they don't want to worry about their level, or they don't want others to worry about their level, and that's a signal that's not available to us.

[rrenaud]

We could take the official leaderboard and drop names that don't appear on it.

I used past tense with samojedi because I played against him like 6 months ago, but haven't seen him since.

[Fabian]

Well, finding myself in second place on a leaderboard was a nice surprise!

[fellowmartian]

Hmmm. On old iso I was 35, new iso 31, new Councilroom 28...

I'm confused.

[rrenaud]

The many clocks problem .  It's why I didn't want to put up a leaderboard on cr unless it was compellingly different.

Levels are very sensitive to implementation details in the scoring system, even to the degree of picking different constants, leaving aside things like whether or not to increase players variance over time.  Ranking systems are also unconstrained with regards to scale, you could multiple every ones level/variance by 2 and the ranking system would be doing it's just as well.

Ranks (or even rank / #players for compensating for systems that filter out certain players) should be a more consistent measure across different scoring systems.  Note that multiplying everyones level by a (positive) constant doesn't change the ranking.

Further, you might even look at the variance in players ranks across different score boards to see how much players successfully gamed the rankings. 

[fellowmartian]

Okay, I've just seen that yaron is level 24.

That's ridiculous.

That kind of thing completely blows out of the water the idea of getting a reasonable match-up based on level. Yaron will pummel your 'average' level 24.

I'm sure everyone here is massively grateful to dougz for isotropic. This is one small thing compared to all the unbelievably fantastically good things he's done to make isotropic what it is, so I hope all the moaning about this doesn't come across as getting on his back.

But.

Dougz, are you there? Can you see all the people above with very good reasoning behind why this change is so so much for the worse? If you agree, can you change it back, please? If not, can you tell us why?

[theory]

I'm curious about something else:

On Nov 27, I was level 26;
on Nov 28 I was level 25;
and today, Nov 30, I am level 24.

I would have assumed this was because my old games are going away, but my # of eligible games is the same at 129, and I have not logged into Isotropic or played a game since the new leaderboard.  What gives?  Why do people's rank slip even when the number of their eligible games stays the same?

[cherdano]

I really like the new rankings, since I am now 10 levels better than theory.
Seriously,  there is something wrong.

[toaster]

It appears that dougz doesn't post here very often in general...theory or rrenaud, I presume that you've had some contact with him before, would either of you be willing to email him and see if he can provide an update on this issue?

[WanderingWinder]

I'm curious about something else:

On Nov 27, I was level 26;
on Nov 28 I was level 25;
and today, Nov 30, I am level 24.

I would have assumed this was because my old games are going away, but my # of eligible games is the same at 129, and I have not logged into Isotropic or played a game since the new leaderboard.  What gives?  Why do people's rank slip even when the number of their eligible games stays the same?

If your rank is based only on the last 30 days, your 'strength of schedule' has got to be changing. Because if you count your games for the change they bear on you based on how strong your opponents were ranked when you played them, then that carries the information of how they were ranked then into your rating, which is information from before that 30 day period.
And yeah, I know that that doesn't really make sense.

[olneyce]

If your rank is based only on the last 30 days, your 'strength of schedule' has got to be changing. Because if you count your games for the change they bear on you based on how strong your opponents were ranked when you played them, then that carries the information of how they were ranked then into your rating, which is information from before that 30 day period.
And yeah, I know that that doesn't really make sense.

I thought the system merely calculated the current ratings of people when they play, and adds or subtracts at that moment.  I didn't think it was retrospective in the way that you're describing.  Is that wrong?

As in: people who beat me like a drum when I first started playing don't get credit for beating a high-ranking player, since I was a level 0 at the time the match actually took place.   Right?

[WanderingWinder]

If your rank is based only on the last 30 days, your 'strength of schedule' has got to be changing. Because if you count your games for the change they bear on you based on how strong your opponents were ranked when you played them, then that carries the information of how they were ranked then into your rating, which is information from before that 30 day period.
And yeah, I know that that doesn't really make sense.

I thought the system merely calculated the current ratings of people when they play, and adds or subtracts at that moment.  I didn't think it was retrospective in the way that you're describing.  Is that wrong?

As in: people who beat me like a drum when I first started playing don't get credit for beating a high-ranking player, since I was a level 0 at the time the match actually took place.   Right?

That's definitely the way it WAS. Now... I actually have no clue what it's doing now, since it doesn't make sense to me how you can really do a pool based on 'the last 30 days', where those 30 days are constantly floating.

[Kirian]

I'm curious about something else:

On Nov 27, I was level 26;
on Nov 28 I was level 25;
and today, Nov 30, I am level 24.

I would have assumed this was because my old games are going away, but my # of eligible games is the same at 129, and I have not logged into Isotropic or played a game since the new leaderboard.  What gives?  Why do people's rank slip even when the number of their eligible games stays the same?

If your rank is based only on the last 30 days, your 'strength of schedule' has got to be changing. Because if you count your games for the change they bear on you based on how strong your opponents were ranked when you played them, then that carries the information of how they were ranked then into your rating, which is information from before that 30 day period.
And yeah, I know that that doesn't really make sense.

Yeah, that probably makes sense.  Presumably what happened is that theory played no games between 27 October and 30 October, so the number of eligible games wouldn't have changed.  But the ratings of the people against whom he played would now have also changed... and I guess would have changed at the time he had played them?

It seems like the new system is basically setting everyone to 25 +- 25, then running through the past thirty days of logs and parsing forward from there.  I trued to type the following in just a couple of sentences, and realized English doesn't have the verb tenses to deal with it, because it's basically time travel.  Let me try this way:

On 27 Nov, rankings were calculated based on 27 Oct to 26 Nov.  Theory played no games in the first four of those days, but other players did.  At the time theory played his first games, those players had ranks, presumably better than 25+-25, based on games they played between 27 Oct and 30 Oct.

On 30 Nov, rankings were calculated based on 30 Oct to 29 Nov.  The first players theory played now, instead, were all 25+-25, so theory's rank increased less for those first games.

----

In other words, if I play 100 games today and then do nothing for 30 days, during that time my ranking will regress to the mean (25+-25) based entirely on the rankings of other players shifting as the time window used to determine peoples' rankings changes.

In other other words, this is a bad time travel novel plot.  A computer program is erasing things in the past, which is causing changes in the present.  All we need is an intrepid team of time travellers (I'm imaging theory as the scientist, rrenaud as the grizzled space marine, guided as the hardened veteran, myself as the redshirt who dies in a spacetime anomaly, and of course Sam Rockwell as the Plucky Comic Relief) to destroy the computer--at some point in the past, before it starts erasing said past.  Or something like that.

[rspeer]

Kirian, that is an amazing description of what's going on.

I think I might have gotten dougz to read this thread by e-mailing him earlier today. But he sounded unconvinced. I fear I did so too early, because he needs to see your post.

[ackack]

I'm unconvinced there's really a problem; at any rate, I'm not convinced it's nearly as bad as everybody thinks it is. A couple of thoughts:

- Everybody seems to be assuming that the leaderboard before was basically correct. Certainly it did a pretty good approximate job - top leaderboard players were legitimately good. But top players now are still legitimately good! There seems to be a substantial status quo bias here where the more recognized names are assumed to have been better, but I don't see a ton of evidence put forward to necessarily support this. It's just an assumption.

- Among other things, something to consider is that the old ratings incorporated a lot of data from a different era of play, where things like only playing Colony games was still allowed. The new 30 day approach presumably does not reflect that data at all, and thus has at least some chance of being more accurate in the given climate.

- The best argument for what's weird is that undoubtedly good players who go inactive for a while can have an unusually low level. But if it's strictly the level, how important is that? Having a leaderboard based on level as opposed to estimated skill (another way in which the initial leaderboard might be a bit misleading, although this adjustment would not have produced a huge shakeup) seems to me to lead to some weird distortions in that way. Playing somebody who has fallen down the leaderboard because their random deviation is now 15 instead of 7 does not necessarily mean that your TrueSkill will be unduly punished.

- Resetting to complete unknown status does seem wrong. Here are a few ideas to fix that main problem: ratings floors; mean skills that persist through the 30 day window, which still determines the random deviation; reverting to the old way of doing things, but only taking data from the post-Paralyzed change where restricted games do not contribute to leaderboard.

[theory]

- The best argument for what's weird is that undoubtedly good players who go inactive for a while can have an unusually low level. But if it's strictly the level, how important is that? Having a leaderboard based on level as opposed to estimated skill (another way in which the initial leaderboard might be a bit misleading, although this adjustment would not have produced a huge shakeup) seems to me to lead to some weird distortions in that way. Playing somebody who has fallen down the leaderboard because their random deviation is now 15 instead of 7 does not necessarily mean that your TrueSkill will be unduly punished.

I'm actually OK with the leaderboard, since it does accomplish the desired purpose of having a more transient "what's hot" board instead of an all-time thing.  (I prefer a leaderboard that's all-time, but I can accept the alternate as well.)  But it's not just level that's changed, it's rank as well.  This means that:

1) Automatching based on level (the original purpose of TrueSkill) is now kind of pointless/misleading;
2) You introduce a lot of inconsistent junk data into the system, since you have people who get underrated because they're playing against underrated players, and that in turn underrates the people that play them.

[DStu]

But it's also that the system thinks it is much more sure about the skill. Here the variances from before and now:

Axis: x:number of games played; y variance in TS.

Note that the new ones are almost determined by the number of games you have played, they are all near c/x^2.

[Reyk]

@DStu: Would you mind to explain your post to Joe Public? ;-)

[DStu]

@DStu: Would you mind to explain your post to Joe Public? ;-)

No.
I took the leaderboard from the two dates. Took the number of played games (x-axis) and the Variances in the skill (y-axis) and plotted them.

You see that now, the variance is lower than before by the same number of played games. Which per se is not surprising, as most of the games are cut of now. But even if you consider this, the guys with ~>500 games this month have smaller variances now than everybody on the old leaderboard. So somehow I have the impression that some more parameters in the TS changed, causing it to get confident faster than before.

The second thing is that this confidence does not depend on anything anymore but the number of games. It is more or less on the line 1/(games)^(1/2), so the variance decreases with the squareroot of the games. More or less independent of how you play, if you lose/win most of your games or if you win half of them. Before, that was defintely not the case.

Edit: Just the new variances:


Edit2: And it should not be the case in TS that the variance just decreases by the number of games, because other parameters also enter: What was the skill diference to your opponents, what was their variance, etc...

[rod-]

Before:  Trueskill variance (the number subtracted from your calculated skill in order to generate your level) was loosely correlated with the number of games you played, particularly if you had played less than 500 games.

Now: Trueskill variance is almost exactly correlated with the number of games you have played, no matter how many games you have played. 

The difference here is of course because before, you could accumulate 1000 games in many different ways: either by playing 100 games on 10 individual days or 1 game on 1000 individual days.  The way the system was set up, the person who played 100*10 would have a much lower variance than the person who played 1*1000, because your variance would not increase if you did not play on any given day.  Now, you have 30 days.  Full-stop.  If you don't play on 1 of those 30 days, your overall variance will still change.

I sort of like the change, because I no longer feel obligated to "only" play when i have time to play 7+ games.

 That does not mean i endorse the new variances being so much "lower" than the old ones.

[DStu]

I sort of like the change, because I no longer feel obligated to "only" play when i have time to play 7+ games.  I can't hypothesize any logical reason why the variance should not be a straight function of the number of games played.

The reason is that Trueskill's variance is not only a function of the number of games:

The uncertainty of both players (regardless of win/loss/draw) is going to decrease by the factor 1-σ2/c2 * w((μ_winner-μ_loser)/c,ε/c). Again, the player with the larger uncertainty gets the bigger decrease.

http://research.microsoft.com/en-us/projects/trueskill/details.aspx  (also for what the functions mean)

If you play win against an opponent that is ranked (much*) lower than you, your uncertainity should not decrease. Same if you lose against somehow that is (much*) higher than you. [* what much means depends on the parameters].
What is happening now I think is the following. If at date now-30days, everybody is 25+-25, than for a long part of the intervall we consider, nobody is "much" higher than anybody else, because everybody is 25. So no matter if you win against someone who is 5 or 45 in the end, at the beginning he is 25, you are 25 and the uncertainity decreases the same, no matter against whom you win or lose.

[Reyk]

I sort of like the change, because I no longer feel obligated to "only" play when i have time to play 7+ games.

It ist true that you could lose a level under the old system by playing one game per day and winning it.
But now it's much much worse. You simply can't use automatch +/- rank as theory already stated.

[rod-]

I suppose it's easy for me to gloss over that "loss", as I've never used automatch +/- rank.  I'd rather get a fast match against a level 7 than wait for a level 37.  However, I will also note that under the current system, I can't honestly tell a difference in the level of play between the 7 and the 37.

[ckb]

- The best argument for what's weird is that undoubtedly good players who go inactive for a while can have an unusually low level. But if it's strictly the level, how important is that? Having a leaderboard based on level as opposed to estimated skill (another way in which the initial leaderboard might be a bit misleading, although this adjustment would not have produced a huge shakeup) seems to me to lead to some weird distortions in that way. Playing somebody who has fallen down the leaderboard because their random deviation is now 15 instead of 7 does not necessarily mean that your TrueSkill will be unduly punished.

The problem, as far as I can tell, is that it's not just the "level" that is affected by going inactive, it's the skill estimate as well. In other words, as someone posted above, you start out at a skill of 25 as of thirty days ago and then only what you have done since then counts toward your skill.

So, if a formerly-top player happened to only play one game in the past 30 days and lost, his skill would likely be somewhere around 23. This tells you nothing about his true overall level of play. Someone who plays against this "level 0" player will be unfairly punished for losing to this "lower-skill" player.

[Axxle]

I suppose it's easy for me to gloss over that "loss", as I've never used automatch +/- rank.  I'd rather get a fast match against a level 7 than wait for a level 37.  However, I will also note that under the current system, I can't honestly tell a difference in the level of play between the 7 and the 37.

If I can't tell a player's actual skill level from his skill estimate or rank, then what's the point?  It seems all this change caters to is the people who care about rising up the ranks, but not actually caring if their skill level is improving.  I'm still having fun playing, but I no longer have that competitive feeling that the leaderboard used to provide.  That disappoints me.

[greatexpectations]

it's late in the game, but maybe theory or rrenaud could throw a poll onto this thread?  seeing some rough numbers might better inform and/or influence doug. 

and for those that are fine with the new style, they can have their voice heard without having having to face any comment backlash.

[theory]

I suppose it's easy for me to gloss over that "loss", as I've never used automatch +/- rank.  I'd rather get a fast match against a level 7 than wait for a level 37.  However, I will also note that under the current system, I can't honestly tell a difference in the level of play between the 7 and the 37.

If I can't tell a player's actual skill level from his skill estimate or rank, then what's the point?  It seems all this change caters to is the people who care about rising up the ranks, but not actually caring if their skill level is improving.  I'm still having fun playing, but I no longer have that competitive feeling that the leaderboard used to provide.  That disappoints me.

You raise a good point: the leaderboard is no longer capable of tracking a player's skill over time, so the old idea of "improving my game" and "making it to level 40" is now meaningless.

it's late in the game, but maybe theory or rrenaud could throw a poll onto this thread?  seeing some rough numbers might better inform and/or influence doug. 

and for those that are fine with the new style, they can have their voice heard without having having to face any comment backlash.

Done.  See separate topic.