https://www.teamfortress.tv/55380/post-season-survey-rgl-6s-s2

I think its important to maintain the current current system, rather than only consider a team's win/loss ratio. What is the purpose of the regular season really? To seed for playoffs by determining the best teams of each division, with individual matches as the only metric for doing this. Having points that show a "dominance score" per match (how much a team won by) will ideally show how much "better" one team was than another, in addition to incentivizing better teams to not throw against worse teams, which is a problem I have seen many times before. Given this dominance score, against worse teams, each better team will have an opportunity to show how much better they are than said team, by the round difference translated into point difference at the end. Essentially, the point system currently in place actually provides more data to the league of how skilled teams are relative to each other rather than a simple win or loss. After all, compare a 5-4 victory to a 5-0 victory. The team that 5-0's is clearly much better than the team that gets 5-0'd whereas a 5-4 shows a hard fought and close match, indicating the teams are close together in skill level, and that each team should be deserving of points to determine seeding at the end of the season.

tldr more data > less data when it comes to the league ranking teams for playoffs and statistically its not always correct that a team with a better w/l is a better team overall

^no

teams shouldnt be penalized for winning a close match

in a situation where every team plays every other team there is no reason to not use win/loss
if next season that isnt the case, then maybe point would make more sense

conversely, a team that loses a match 5-4 should get no compensation at all? again consider the purpose of the regular season should be to measure the skill level of teams as accurately as possible in order to have the absolute best teams in playoffs.

ok you obviously didnt even read my post

or i just disagree with you? who needs a "dominance" statistic when dominance is already demonstrated by their w/l compared to the rest of the div, and if teams are TRULY close in skill theyll end with similar records so match points are worthless anyways. the match point system is just pity points, win the fucking match and prove youre at the same level

nobody wants JUST win loss, people want win loss to be the deciding factor, then have the points system be used for tie breaking.

i understand, but i'm just confused as to why people want w/l to be the deciding factor when points are more representative of teams' relative skill. the most accurate metric should be used first for absolute placements, then a less accurate one to tiebreak.

youre assuming people agree with you that points are more representative than win/loss

points are not more accurate than win/loss

how are they not? point system already favors the winner, as winning a match guarantees you 2 out of the 3 possible points for the match. the remaining point measures how much you won by, with the other team being able to earn some of it. if you win 5-0 the points are awarded 3-0, as should be deserved by a shutout. this season points mostly lined up with w/l except in the case of invite where h5 demoman didnt make playoffs despite having a better w/l. but if u look at h2h between h5 demoman and ford, their 2 matches were 3-4 in favor of h5, then 5-2 in favor of ford, which would indicate the point system worked for the correct team as the round total across these two matches is 8-6 in favor of ford.

https://i.imgur.com/vqWZSLk.png

here you can see that ford dropped(pd) 2 total points on wins compared to h5's 4, meaning that when ford won a match, they won matches more convincingly than h5. looking at points earned(pe) ford's losses were much closer than h5's which allowed them to earn more points on loss. an overall, generalized, way to put this is when ford loses, they just barely lose, and when they win, they win fairly convincingly. compare this to h5 where the opposite is true: when they lose its by a larger margin and when they win its not by as big of a margin.

the real question is whether you care about how close matches are in addition to concrete result, and personally i think the score itself does contribute enough to warrant the point system. this is all im really gonna say tho ive said everything i want to and people are free to disagree.

the point system is a good replacement for rounds for/rounds against, not a good replacement for W/L

While I think the current points system is a decent attempt at trying to rank teams more accurately or allow for teams that had some unfortunate matches to make playoffs, I don't think that a system where a 16-0 team could theoretically be lower seeded than a 9-7 team is a good one (this can happen if the 16-0 team won every match 5-4 (32 points), while the 9-7 team won all their matches 5-0 and lost all their matches 5-4 (34 points)). I think that if a team wins all their matches, they deserve to have the top seed (while this situation is really unlikely, I think that this extreme shows some of the issues with the current system). Also, a team's ability to be able to close out/win a match is, in my opinion, an important one that is more emphasized in win/loss rather than the current points system. If there were some changes to the points system that rewarded winning more, I think I would like it, but for now I think that using win/loss as the main ranking system and points as a tiebreaker system is better.

What if the points system always gave 3 points to the winning team, and up to some amount of points (my example below gives a max of 1.2, but this can be changed) for the losing team getting rounds? I think the big flaw of the points system is that a team can get 5 rounds, win, but receive different amounts of points for that win, which stems from the "every match gives out exactly 3 points" mentality, which I think is dumb because not all matches are the same. What if, let's say winning got you 3 points, and rounds on loss got you 0.3 points? So if team A beats team B 5-3, then team A get's 3 points for winning, and team B gets 0.9 points for at least putting up a fight and capping rounds. This would also mean that winning guarantees you 3 points, and it's still worth it to get rounds if you lose, but the spread in points you get from winning and losing is larger than it is now so the "better w/l has less points" is less likely. Depending on how much you value rounds won while losing, you can change the value of how much each of those rounds gets you.

I'm also not against using straight w/l for placement and points or some similar system for tiebreakers, however the points system can be changed to more heavily favor winning.

tl:dr turn 2-1 and 2.5-0.5 win scores into 3-1 and 3-0.5 scores to reward winning more

I just don't see how in a competitive game, you can look at like a 8-8 team and say "yeah their eight losses were closer than those of the 13-3 team so they deserve to be ranked higher" despite losing way more games. Like I understand when the W/L is tied or there's only a one win difference, but when the records can be that far apart and result in such wacky standings, that's the main problem.

As air said, having points come after W/L is perfect. But you play to win the game. Teams that win should be placed higher than the teams that didn't (as much).

Also I don't care how "good" a team claims to be, if they don't win enough games to get into playoffs then they don't deserve to be there. Speaking of, the fact they let half of all non-Invite divisions in playoffs is another problem, but that's for another time.

This is an interesting idea but the thing is with the current system there are only so many points to be gotten out of a match for both teams combined, which balances things out. This is good because every match can only shift the total points in the system by 3, whereas with your suggested system a 5-4 match would shift the overall rankings by 4. I mean if anything this goes against the idea that you reward dominance, because if two teams want to screw another team over their best bet is to just have close games and get as many points for each other (e.g., 5-4 both ways yields 4 points for both team total).

You don't want close matches to be disproportionately good for both teams, I can see potentially rewarding getting to golden cap like this and giving a pity point or something, but even still it just throws a wrench in the whole system. If you're going to reward teams for getting rounds, you should penalize them from dropping rounds. The solution I see is simply adjusting it so that the wins matter more, which there are many ways to do such as making each round count for 1/6th of a point instead, so if you lose 5-4 you still get 0.66 points (instead of 1).

As everyone else has said W/L and head to head are definitely better metrics for seeding, but points as a third way to break ties makes sense. If you start off poorly and choke at the beginning and fix your stuff in the second half and come back second half, the only thing that matters at the end of the day is the very last round if you won or not (I mean you're not going to say that the pats coming back from being down 28-3 is less of a win lol)

Also I think the pick/ban system for the regular system should be rethought (but maybe other people disagree: I made a twitter poll out of curiosity). The concept of home/away is interesting I guess but it seems to give the home team way more power than they should have imo. Currently the home team bans 2 maps, then the away team bans 2, then the home team picks from the remaining 3.

In theory this seems like a decent idea, and it's very similar to the ESEA system (3 bans each independently, then randomized). The key difference is that in the majority of cases, the away team simply bans the two maps they want to play least, and then the home team effectively gets to pick from the remaining 6 maps.

If the home team happens to read the other team poorly and bans maps that the other team was going to ban (say, via and granary), despite knowing that they will never have to pick those maps if they don't want to, it can be the case that the away team gets to do two meaningful bans (say, banning snake and gully which the home team is better than them at). But even with this, the home team wasn't going to choose those maps anyways, the home team can always effectively choose between 6 of the maps which they want to play.

ESEA's system was interesting because you didn't know what maps the other team was going to ban, you could risk leaving a chaos map (clearcut, via, gran, etc) in if you think the other team will ban it, but then you run the risk of them preferring to ban a map you are really good at and then you get RNG'd into the map neither team is very good at. This gave good teams incentives to actually play the new maps so they can use their bans meaningfully and avoid playing the maps that their opponents were best at, instead of avoiding the 'unfun' new map which in reality might not be that hard to figure out (e.g., ascent practiced propaganda and was solid at it).

edit: given all this stuff to talk about I think there should definitely be either a specific invite survey or a meeting like the LAN one, there's a ton of variables to look into as well as this such as how many maps. The ESEA system had 9 maps which I think worked better but maybe they prefer to keep the number of maps consistent with other divisions (understandably). Usually the 3 maps would have some overlap anyways so having 3 bans out of 8 isn't too crazy, and even if they don't overlap, randomly choosing between the two remaining maps that both teams want seems fair enough.

head to head record is a better way of breaking a tie in the first place imo the point system just seems silly

How about having a set amount go to the winner and have the remaining get divided up by rounds? Or maybe use difference in rounds so teams can't run up scores, a 5-4 would give same points as 1-0.

regardless of what you think you should really rethink your brick wall theory when discussing what would be best

doesn't matter what YOU think because YOU don't speak for everyone

imo, a vote would be a much better compromise because ultimately what the people want should be the deciding factor in how a league is played

it is the endless back and forth with no compromising that will eventually divide a dying game (CEVO, FACEIT)

but if u ask me man that point shit is dumb as hell

This is how the system works currently. 2 of the match points go to the winner, 1 gets divided up by rounds.

This is a flaw that was realized this past season, but would be fixed for next season if the point system is kept.

The only scenario in which a point system works is if the winner of the match does not suffer from losing points if the match is close. It should be fine if only the loser gets awarded partial points accordingly (and the winner gets the full 3 points).