We need fewer divs in Highlander

#1767
Topic created · 20 Posts · 234 Views
  • Something worth noting is that when UGC added Gold division and split steel into red/blue there were 300 teams playing highlander. They removed Gold when there was what, 80 total teams? RGL only has 60 teams now, so I believe there is valid argument for consolidating lower divisions.

    Editing my post with more accurate data:
    http://imgur.com/9Rrgrxr this chart shows the total number of live (played at least 6 weeks) teams per season of ugc, for reference Gold was added in s9 and removed after… s21? I might need a fact check if it matters

    Total number of ugc hl teams vs esea 6s teams by season is here:
    http://imgur.com/17hMH2E

  • I agree that it would be good to deemphasize “moving up divs” as a seasonal goal. People fixate on it excessively. However, I don’t see how significantly increasing div size would improve the efficacy of the Swiss structure beyond a certain point, as eventually it just introduces too many opportunities for meaningless ties, and the fact that it takes time for teams to settle into the right area of the rankings means that a larger div = a larger number of meaningless rolls every season (whether doing the rolling or being rolled), compared to a small, curated RR div. Rolls teach nothing to the winning team and are suboptimal learning experiences for the losing team. It seems to court calcification rather than improvement.

    In the end, flexibility that takes into account team numbers and skill distributions from season to season is always important, but if there is an issue of divs being too small for Swiss to work, I don’t think “just smash all the divs together” is necessarily the best or only answer.

  • It would be great to get rid of the mentality of div climbing for glory and status, and incentivize personal and general team improvement. As a bonus it will also practically remove sandbagging as a thing. It’s much more fun to keep a team together and improve together than to switch teams every season just to try to “be on a better team.” However, I believe having 3 skill based divisions will cause too large of a skill gap between the bottom and top of the divs. Like what @scaredy-bat said, the number of rolls that will occur in a regular season will increase drastically and may even be a turn off for some players; causing them to give up and quit because of being badly beat repeatedly instead of encouraging them to improve. More teams in divisions will also make officials more luck based; if you are a mid skill level team and keep facing lower skill level teams, you may earn more points than another team that is more skillful than yours simply because of luck. I think removing challenger is a good move, and perhaps even removing main later down the road, but 3 divs seems too extreme.

  • @scaredy-bat said in We need fewer divs in Highlander:

    However, I don’t see how significantly increasing div size would improve the efficacy of the Swiss structure beyond a certain point, as eventually it just introduces too many opportunities for meaningless ties, and the fact that it takes time for teams to settle into the right area of the rankings means that a larger div = a larger number of meaningless rolls every season (whether doing the rolling or being rolled), compared to a small, curated RR div. Rolls teach nothing to the winning team and are suboptimal learning experiences for the losing team. It seems to court calcification rather than improvement.

    @Discount said in We need fewer divs in Highlander:

    However, I believe having 3 skill based divisions will cause too large of a skill gap between the bottom and top of the divs. Like what @scaredy-bat said, the number of rolls that will occur in a regular season will increase drastically and may even be a turn off for some players; causing them to give up and quit because of being badly beat repeatedly instead of encouraging them to improve.

    Rolls are useful in a Swiss system as intended with large divisions: to set up the early structure of the division. The point is that similar skill teams should face each other to differentiate: which is the better among them.

    Rolls will be concentrated towards the beginning of the season, when there is less seeding to go off. This is natural, inherent to Swiss, and most critically, happens anyway. Towards the beginning of each season you’re far more likely to see teams with large gaps in skill face each other. Yes, you’ll see harder rolls, but let’s face it, there’s very little meaningful difference between a 3:10 offensive time on Vigil, a 3:40, and a 4:10 time - which already happens, and happens a lot. There’s very little difference between being held on cliff/grass on Product and being held on point if either way, you can’t get any cap time the whole match.

    But what happens right now that won’t happen in my proposed world is that, once teams have exhausted all the teams near them in skill, they start hitting better or worse teams, to the point of it being a roll, just because there are no longer any teams in their division close to them in skill.

    This is most apparent right now in the teams that roll their divisions, with maybe only one or two worthy challengers. Once these teams have played said challengers, they are now facing teams they will reliably roll and that have been paired with said top team just out of poor luck, since matchmaking is done from the top down and the top team has already played everyone in between. Meanwhile, they often have worthy opponents in the div above - and often would beat, but not roll, the bottom teams in the div above. These matches would tell us so much more about their skill and where they deserve to be than just “at the top of division X” right now. They would tell us more about where they deserve to be placed next season. They would tell prospective tryouts more about what they can expect from that team and its players in the future. Sure, at the top of the divs there would still be the possibility of teams rolling the div, but fewer divs would reduce the frequency of those teams just because there are fewer divs to roll, and fewer divs would make it easier to have clear cut boundaries and overall streamline div placement for admins, reducing the chance of unsuitable team placements.

    In the other scenario are teams near the bottom of their divs, especially the ones who have unusually hard seasons. These teams have such hard seasons because of poor luck; fewer teams in each div makes for more chance that such poor pairings happen. The best example I can think of this happening is Sunrise from this past season of HL. They faced five out of six playoffs teams - 3-0ing one of them - and facing only two non-playoffs teams (and soundly beating both) - yet placed last seed at the end of the regular season. Just because there are fewer imbalanced matchups in a Swiss system with more teams, there is less chance for matchups to be exhausted to the point where a team gets screwed by the system and not by just being less skilled.

  • @Discount said in We need fewer divs in Highlander:

    More teams in divisions will also make officials more luck based; if you are a mid skill level team and keep facing lower skill level teams, you may earn more points than another team that is more skillful than yours simply because of luck.

    To that end, this just happens less when Swiss is working as it is intended to, with lots of teams. You are put up against teams closest to you in the scoreboard (i.e. with near equal match points) and you don’t end up facing consistently lower (or higher) skill teams since you’ve exhausted all the teams of similar skill level.

  • I’ve said it to a few different people recently, that I believe TF2 competitive would benefit from an open-ladder system (maybe a dual ladder system?), longer seasons, more flexibility in match times (less reliance on traditional Monday matches, weekend scrims, etc.), more monetary incentives (every team has to pay a flat fee to enter), and much more. I have even toyed with the idea of having it be a challenge system, where teams can play (x) amount of matches a week (if they please) to earn points throughout (x) amount of months, where the playoffs bracket is dictated by those who put in more time / effort / energy, and did what they needed to get to the playoffs. I have many different ideas for ways TF2 competitive could evolve (in my opinion it would be evolution).

  • @civ I have a similar mindset honestly. One thing that perplexes me is why the 6’s season is two weeks longer than HL. At the very least, I am hoping that we can run HL for 9 weeks just like 6’s, which would open up two additional invite slots, making challenger completely useless (even if they already removed it).

  • @civ

    more flexibility in match times (less reliance on traditional Monday matches, weekend scrims, etc.)

    One of the biggest things holding back HL currently is really just this. A lot of 6s players (top 6s players for that matter) would be able to play HL if it didn’t conflict with their 6s schedules. This is because Tuesdays are normally match days for IM+ in 6s, and Mondays are used as scrim days in 6s. Unfortunately, Monday is also used for HL match days, and with HL matches taking long to start, it causes an overlap.

    There are also different suggestions that you mentioned such as more monetary incentives (league fees for other HL divisions that are not Invite), more matches (would depend on if HL players are willing to deviate from this Monday structure), and longer seasons which could make Highlander taken more seriously.

    These things have been polled and whether the HL community wants to take the next step is open for discussion.

  • @Melon tbh surprising to see how much smaller esea was though tbf a lot of weaker players were in ugc 6s and esea was a paid league

  • this is only tangentially related to div size but i didn’t really realize that until i had typed up most of this because it’s been on my mind for a while

    i don’t have the knowhow for the methodology to prove this statistically and i wouldn’t really have the drive to do so even if i did, so you’re gonna have to trust me when i say that a significant amount of frustration re: swiss comes from the fact that RGL runs an odd number of matches

    anecdotally speaking, six weeks is generally a decent amount of time for the general structure of a division to form given the average size of divisions. further, due to how the bottom of playoffs tends to be around the 4-3 mark in the higher divs, it means the large mass of 3-3 teams swiss helps to guarantee in week 6 have an inordinately large effect on playoff seedings, because of all those 3-3 teams, some of those should be 3-4 and some should 4-3 but 6 weeks isn’t enough time for the stratification of those teams by MP to be accurate (even if the general “shape” of the div is). so the top 3-3 team playing “up” a record gets screwed, the bottom 3-3 team playing “down” gets boosted, and within the inter-3-3 matches you could have two theoretical 4-3 teams playing and two theoretical 3-4 teams playing, leading to some cases where worse teams end up making short term gains over better teams. an 8th week would correct a lot this.

    i don’t mean to say that 6 weeks is better than 7 weeks, as each week helps to further align the standings to their “true” form, but if the marginal gain in accuracy from week 6>7 is x, the marginal from week 7 to week 8 would probably be closer to 2x. that would continue further with 8>9 being something like .5x and 9>10 being around x.

    more in line with the topic of the thread, increasing div size to the point that the playoff cutoff is less near the middle of the pack and closer to the top (to the point that lower 5-2s miss playoffs) you eliminate a lot of this frustration, but i’d still be in favor of increasing season length to 8 weeks for swiss divs regardless

  • @civ @exa

    totally agree with both of y’all but that reform isn’t omnibus, you can improve the div structure without all of these things and it takes significantly less political capital to do that than all these other things like league fees, more matches, longer seasons

  • @Alto it makes sense to me.

    it would make sense imo to have, especially in conjunction with the other stuff i’ve suggested, a 10 team RR invite, swiss IM/AM/NC, and 8 weeks/9 for Invite as in 6s. Still one match a week unlike 6s but two matches would be insanely hard to push

Log in to reply