Seeding Standard

[Banks]

In New England we have some people who are unsatisfied with the seeding and who they are put in a pool with or seeded under. We do run pools quite often, but it is always similar first/second seeds. The problem that arises is people who are salty over getting seeded below someone and getting another good player in their pool, thus stopping them from getting first seed in the bracket and potentially getting eliminated earlier than they would have if seeding was slightly changed. This is especially an issue when tristate comes to New England because there are more "guaranteed" first seeds, and more picking and choosing who is 1st or 2nd.

The TO proposed basing seeding strictly off the previous tournament in the series, but that could obviously be quite inaccurate depending on who attends or if someone causes an upset in the bracket. I suggested maybe basing it on the last three tournaments or something, if anything.

Anyway, I suppose this is not really a huge problem that couldn't be solved by saying 'deal with it' and moving on. I was just wondering what most people base seeding on for long running regional tournaments, and how much weight they put on different criteria that could be argued by someone to get a higher seed.

 

[Marc]

I usually:
-Have as many top and high seeds as there are pools, often even third seeds.
-Have TIO randomly assign them to pools.
-Separate people from the same region, as well as team partners (though usually people team with someone their own level and are not that likely to be in the same pool).
-Have other people check the pools before the tournament start to catch potential problems.
-I sometimes do change pools or parts of the bracket that seem particularly stacked, but not if I can avoid it.

I take my time with it and we usually don't have drama over seeding. In what scenarios do people ***** about what seed they get for pools? You typically only have 2/3 seeding groups, so that means you have high seeds asking for top seed? If it's that gray area I'd say to just go by most recent results, but at some point people will also just have to accept that they need to prove themselves more and that the TO's decision is final.

 

[Banks]

If there are enough pools for each of the group of top players to get their own pool then it works out, but in many instances two of those people will have to play each other. The reaction is something along the lines of "Why is he in my pool, why didn't you put him in ____'s pool? Why is he first seed?!" I don't think people should even be too concerned about it for pools at a regional, but sometimes it happens.

Thinking back, problems have arised after pools when making bracket as well, because in general there are a bunch of first seeds who will be the same at every event. Random brackets usually create a bunch of conflicts and crews/teams etc have to be separated, and sometimes it creates situations just because of the people there.

An example is when th0rn beat Darc at one tournament even though he hasn't beaten him before, and they have played multiple times at different venues/events. At the next tournament th0rn was seeded above Darc, and although I don't think he cared at first, he wasn't too happy with the bracket.

It wasn't a huge deal in terms of people raging or anything, but was something that basically made you choose between recent results and long term results. Since the idea to base seeding exclusively on the last tournament in the series was thrown out there, I was wondering if that would actually be an improvement or would just piss people off.

 

[Marc]

Well, at the very least it's consistent and something that will be known upfront. You could try doing it and see how it works out in terms of complaining. Separating team/crew mates is reasonable to a certain extent, but if the player pool is limited and the top is fairly solidified, you will see the same matches happen a lot. It's hard for me to judge what will and won't piss people off though, considering this is already more of a problem for you than it ever was for my scene. =/

 

[Zankoku]

If you're doing local seeding, then location doesn't really help, I guess. The most impartial way of doing it would be to implement some sort of ELO system and then seeding pools in this fashion:

Pool 1|Pool 2|Pool 3|Pool 4
1|2|3|4
8|7|6|5
9|10|11|12
16|15|14|13

It'll also eliminate a lot of "why is X player in my pool and not someone else's?"

 

[HyugaRicdeau]

Yeah, Ankoku's seeding is the optimal paradigm; you basically zigzag the seeds through the pools. Then you go through and try to separate people of the same region, trying not to move people too far from their nominal seed. I tend to just seed as far down as it can be done with reasonable precision - I look for a nice cutoff point or tier boundary.

As for how you actually seed people, if you have a recently updated rankings, use that. You should tweak the seeding from the rankings slightly after each tournament, but use the rankings as a sort of tiebreaker. If you don't feel this can be done with any sort of precision, just separate people into tiers with the tio player skill and let it seed randomly.

 

[AlphaZealot]

Actually the most impartial way of seeding is random seeding, which is what the Kishes did at FC and given their three rounds of pools (that get seeded after that based on the previous rounds results) was absolutely the most amazing thing ever.

 

[Marc]

It probably wasn't so awesome for people who traveled there from the same region and good players who randomly had to play each other first round though. At Pound 4 several Europeans (from the same country!) and like all of Puerto Rico were put in the same pool because regional seeding sucked and it was the gayest thing ever.

 

[AlphaZealot]

I said impartial.

And Pound 4 was not randomly seeded first round of pools, at least not for Brawl...

Truly random seeding should have, by statistical improbability alone, spaced people away from each other.

 

[Marc]

I agree it's impartial, but I thought you were implying that it is superior. I'm not sure what rounds of pools this happened in, but it honestly shouldn't happen in any round when you have like 20 foreigners in a pool of 350 players.

Talking Melee btw, though I should also point out that in both games I played the same person in consecutive round of pools (RaynEX in Melee and Mikey Lenetia in Brawl) and was literally told "we don't know how to fix that" when I went and asked about it.

 

[AlphaZealot]

Pound 4 was...not well run bracket side of things (or ruleset side of things, either).

For example, had they used tio properly, it would have been impossible that two people from a previous round of pools would have been placed in the same next round of pools.

(part of me wonders how much bias was likely involved in seeding considering they had to bypass tio auto-seeding)

 

[Marc]

Yeah, that's what I was thinking too. Ultimately I think the best players tend to come out on top unless something is really off, but you want the fairest and most fun experience for everyone. But well, that's in the past.

 

[Zankoku]

I said impartial.

And Pound 4 was not randomly seeded first round of pools, at least not for Brawl...

Truly random seeding should have, by statistical improbability alone, spaced people away from each other.

See, I used to think that, until I went to MLG Columbus 2010.

 

[Banks]

It's hard for me to judge what will and won't piss people off though, considering this is already more of a problem for you than it ever was for my scene. =/

Luckily any issues seem forgotten quite quickly by anyone who does complain, so there's no actual lasting drama. I think it should be interesting to see how it turns out, and it'll be simpler to have concrete facts to fall back on.

If you're doing local seeding, then location doesn't really help, I guess. The most impartial way of doing it would be to implement some sort of ELO system and then seeding pools in this fashion

Yea that's a good way to do it. I think if we end up just looking back one event, then next time we could easily keep going from there in terms of stat tracking. It probably won't take too long to figure out what works, hopefully.



but really tho, if all else fails it's random and people play their siblings first round

 

[AlphaZealot]

See, I used to think that, until I went to MLG Columbus 2010.

There was one MLG event where the seeding, which was suppose to be random, was actually ordered (based alphabetically or when they registered) and the only exception to this were players who already had seeding points. I believe the event was Columbus but at this point I forget. Unfortunately I did not get to structure the brackets except for doubles at any of the MLG events (so the seeding was all done by someone else).

Also IIRC when I check the alphabetical thing was also weird but I forget why. Regardless, I'm fairly certain it was the columbus event, and it was a mistake.

 

[KishPrime]

Actually for FCs, I just grabbed the top 64 (my opinion), split them up by region, and made sure that there were two top different-region players in every initial pool. That covered two of the 4 "advancing" spots, and I don't think I was ever wrong on a single person, while there were still two spots left for everyone else to fight over. I filled up the rest of the spots randomly by region so there was diversity in each pool. Then every pool round after that was, as you say, completely random but seeded.

For smaller events, we just did seeding by crew, for the most part. It usually works out pretty much the same as seeding by skill and it is less controversial.

Also, I like your "deal with it" solution best. TOs need backbone, and that goes a long way. Nothing is ever totally fair, so be up front about it, and if they still complain, tell them to "take the time you spend whining over pools and go practice instead."

 

[AlphaZealot]

I remember being explained by either you or squared years ago that it was completely random seeding, though in retrospect it may have simply referred to round 2/3 of pools (post-tio era, where it did pool seeding for you based on the previous round results).

Also, I hate it when a TO flubs a decision. Make a decision, make it quickly, stick with it. The worst thing I've seen at tournaments are TO's getting bogged down in the decision making process and taking information from too many people.

 

[KishPrime]

Completely random is only guaranteed to completely randomly screw someone over.

It was random, for all practical purposes. Just random within loosely defined subgroups.

 

[Strong Badam]

TOs definitely need backbone. There is always SOMETHING for people to complain about regardless of how well you do; I've had to just go "Deal with it" on several occasions.