An argument for changing the way we classify tournaments
So, tomorrow marks the start of PGR v3’s rollout. It’s a very exciting time for players and spectators alike, especially since this season has been filled with ups and downs for many players. But the upcoming release has me thinking about how tournaments are classified in our community today, and how it could be better. Disclaimer: This is mostly for Smash 4, but it can apply to any Smash game.
How things are now
Currently, tournaments are assessed by four categories — entrants, PGR members, top 20 PGR members, and cash prize pool. Each category has its own weight — for every 64 entrants, a tournament gets 1 point. For every PGR player, it gets 1/4 of a point. For every top 20 PGR player, it receives 2/3 of a point, and for every $2500 in the single prize pool, it receives 1 point. When those points are added up and multiplied by 100, you get the tournament’s tier, from C (lowest) to S (highest).
From my perspective, this is a reasonable way to rate a tournament’s expected quality… But it doesn’t say anything about the actual quality of the tournament. Let me explain this point a little more.
What does it mean to win an A-tier event?
By now, even the occasional spectator knows that Smash 4 is often inconsistent in results. Powerful players may end up at 25th or worse on a given day, as upsets are commonplace. When you win, say, an A-tier event, it could mean that you beat a lot of strong players… or it could mean your bracket was free, thanks to the best players dropping to losers early.
With that in mind, I want to ask — what’s the point of classifying a tournament before it even starts? With a more consistent meta, it might make more sense. It would be almost guaranteed that you have to take out a number of powerful opponents at a large tournament in more consistent games. But what does it really mean to win an A-tier if the top ten lost in pools?
An alternative method of ranking tournaments
Since the current method of classifying tournaments doesn’t show enough, I might suggest a two-tier classification system. The first tier would look like what the PG Stats team has put together, and that would be the potential classification. So for example, pre-CEO, it would have been called a “potential S-tier event.” The second tier would be the actual classification of the event, and describe what happened during the tournament. It might look something like this (EXAMPLE):
For those who don’t know what this is, this is the USFA Event Classification Chart, used to determine the classification of U.S. fencing tournaments. For those who do know what this is, I’ll give you a second to call me crazy before I explain.
…
Okay, the chart. Basically, the ranking process pre-tournament is pretty simple. You need a certain number of entrants who are a certain skill level (A, B, C, D, E) in order to qualify for a particular group. That’s the easy part. The more controversial part is the next part — “Rated Fencers Must Finish.” In order for an event to meet its potential classification, a certain number of the strongest fencers must place in a particular way. Let’s do a quick example.
For a fencing event to potentially qualify as C1, you need 15 fencers, 2 C’s, 2, D’s, 2 E’s. For the actual classification, 2 C-rated fencers and 2 D-rated fencers (or stronger — e.g., A or B) must place within top 8. If a C-rated fencer underperforms that day and ends up in 13th place (leaving only two D’s and one C in top 8), the tournament can only be classified as a D1.
Let’s apply this to Smash: say the top 20 PGR are all classified as A-rated players, and 30–50 are classified as B. If we need 4 A-rated PGR players in top 8 to qualify an event as an S-tier, having no top 20 PGR players in top 8 would lower the event’s classification. This would cause the classification of the event to directly reflect the quality of competition — were the best players at the top of their game? Were there too many upsets? Etc.
But obviously, a system like this isn’t perfect — there are some pitfalls that a Smash-adjusted system must address.
Problems and controversy
Many people in fencing have issues with the USFA classification, due to a couple of reasons. First, it doesn’t really benefit the players who are making upsets. If you’re a hidden legend like Locus, your introduction to the world would be lessened in impact, as the tournament would get a lower classification the earlier you make a splash. Or maybe like ZD, you’re unseeded and win your whole pool. That would be a very bad outcome for everyone involved.
The second issue is Smash-related. Since PGR is released twice a year, players tend to get stronger or fade away in the intermediary period. If you use PGR, a top 20 player in one season may start frequently losing early at the beginning of the next. That means their poor performance will consistently (negatively) affect the events they enter.
Third and finally… no one really wants the performance of someone else to affect how they’re seen. If you win an A-tier because all the top players were eliminated early, you want that A-tier win to last. I believe this is the weakest reason, since the bracket you’re given should affect how you’re seen (and the PG Stats team seems to agree, which is why wins matter in addition to placings for PGR).
Potential solutions
The above issues are up to whoever would want to implement such a system to resolve, but I have a couple of potential solutions I’d like to list really quickly. First, making things right for the players who are upsetting top players — don’t make the number of ranked players required in a certain bracket (e.g. top 8, top 16, top 32, etc.) too large. With this, as long as enough ranked players place well, a few upsets won’t affect the classification. (Though a large number of upsets should definitely lower the classification, in my opinion.)
As for the second issue — PGR players underperforming after being ranked — I might base the classification on seeding as opposed to PGR. Seeding changes more consistently to reflect what has been happening week after week, and so it may be a more flexible predictor of tournament success. Outside of egregious errors in seeding, this method should work fine.
The third issue — the performance of someone else affecting you — is solved by the first suggestion.
Benefits and summary
Despite some details to finalize, I really believe that a two-tiered tournament classification system that takes into account how top players placed would be effective for Smash.
It would allow tournament classifications to be more accurate to what actually happened at the event. There’s no reason for an A-tier where every top player didn’t make it past pools to still be rated an A-tier. While nothing at this level has happened thus far, our current classification system lends itself to the possibility. Events with many upsets — while very interesting to watch — should probably not mean as much if at a certain point, it can qualify as a stacked local.
This will allow for a better understanding of how good a player is, if the classification immediately lets people know how high-quality their win is. Can they hang with the best? Or did they happen to enter a tournament where a top player dropped out last minute, and the rest got upset?
These are all just my opinions, as ever — feel free to criticize in the comments below (and/or on Reddit). I’d love to discuss your thoughts on such a classification system. In addition, let me know what you think about the article itself — too long, too short, too confusing? I appreciate the feedback :)
