Balancing The Roster

[Soldat Du Christ]

Back in january we finaly got a glimpse at how Papa Sakurai determines wether or not to buff/ nerf characters on the now bloated roster. Source:

https://nintendoeverything.com/saku...line-mode-victory-rates-character-usage-more/

Sakurai pays close attention to the Win/ Loss ratios determined by "Elite Smash" match ups, here is an exact quote:

"As for 1-on-1 victory rates, no fighter is lower than 40% and no fighter exceeds 60%. In terms of Elite 1-on-1 matches, the character with the lowest percentage sits at 43.7%, while the character with the highest win rate is at 56.8%. There’s a noticeable difference there but considering that those percentages are for the highest and lowest of 74 fighters, overall you could say that the characters were fairly close to each other."

I'm inclined to assume that this data is gathered internationally, as this would yield the best results following the larger sample size. Now, in principle, this should be the most objective means of approaching roster balance. Not only because of the aforementioned ability of gathering massive sample sizes, but perhaps more importantly; the ability to negate superficial trends. One caviat to this method however is the fact that some vague percentage of this Win/ Loss ratio is a result of the lowest common denominator of Elite Smash matches (matches between less skilled players) where different playstyles and characters are more effective. This is what has led to the phenomena of 'for glory link' or 'elite King K' memes we have become familiar with. Certain characters are more viable against lesser skilled players. This becomes a problem when these results are treated equally to the results gathered at the highest level of elite smash, which i believe is the case.

As players denied of Elite Smash W/ L ratios, we tend to lean on the results that churn out of Tournaments. This method solves some of the issues of Elite Smash as a means of intel gathering. One of which by featuring only the best of the best players in the competetion, which in theory should give us much more clear look at the distribution of power within the roster. But again, this method is also not optimised. Because of the lack of an overwhelming sample size comparable to online play, there is a much greater chance for trends to muddy the waters and create a superficial representation of the percieved distribution of power within the roster.

Both of these methods, while being the most realiable, are still not optimised. Its sad to think that all of this could have been perfected simply by creating a separate online category for only the top 1%ers to compete in, which would provide the most reliable means of gathering data for use in roster balancing. With the recent announcement that there will be no more new modes in smash ultimate i'm afraid this will not possible.

 

[Soldat Du Christ]

BumpBumpBump

 

[Mr.Dale]

I didn't read any of what you said but they should copy and paste ike from smash 4 into ultimate. I mean the new ike is fo BORING compared to the old ike yeah hes better but so is everyone else. Just have his old hitboxes/moves with smash ultimate endlag and we'll be all good.

 

[Predatoria]

Elite Smash represents, I believe, the top 2.5% of players on the roster, or, an Elo rating of around 2k. The problem with creating a separate category for the top 1%ers is, essentially, they've already done this. The percent just is 2.5%, rather than 1%.

Gather a room of 40 Smash players (not 40 at a weekly, or 40 at a tournament, but 40 average people who own Smash Ultimate) and you will have just 1 of those people in Elite if it is a representative sample of the online culture (which is likely, actually, an over-representation of players that are above-average to begin with, as the true average player may not even bother trying online at all).

Now I know compared to people that are extremely good at the game, some of the players in Elite may not seem so tough, but that's because they're well above a 2k Elo rating in skill. If they're 2400, for example, they will win 15 out of every 16 games against a player just entering Elite. They are several standard deviations above the average player and are skilled well beyond a point where the average consumer / purchaser of this game will ever be.

This actually brings up a very important point that I think many people fail to ask on a regular basis.

What skill level should the game developers balance around?

After all, balancing around the top 2.5% of the playerbase (or, as you suggest, the top 1% or less) will have repercussions elsewhere. Do we sacrifice the enjoyment and sense of balance for those other 49 players to create a playing field that only feels fair once you're good enough to play with the pros? Or, do we balance the game around the masses, while letting characters with difficult to execute but highly effective gameplay strategies and mechanics (Olimar, Peach, etc) to really shine at the top. It is not always easy to just "do it all at the same time," or "balance for all levels of play at the same time."

This is the dilemma they're likely dealing with. Yes, some characters are blatantly broken at the top. But, they have to consider how enjoyable their game is for all levels (and types) of play.

My friend goes to a club at uni. It's not a Smash club, but they do play Smash there somewhat regularly, among other games such as Mario Party, Mario Kart, or other common Nintendo party games. All of them have Switches. Most bought smash (bought is a key word for game developers). They queue free for alls with items and stage hazards on. There's several copies of the game sold right there to a group that couldn't care less about the state of the meta. These players see Ganondorf as a god because he hits so hard and can one shot them with his punch that has armor on it. Buff him? The game feels unfair to this group. How do you think High Tier Olimar is going to fare in their hands, in their matches?



Most players want the game balanced around their level of play. Players generally consider designing for balance beneath their level of skill as "catering to casuals," but, unfortunately for people at the top, casuals make up the vast majority of a product's sales. Worse, balancing for only the most exclusive group will end up destabilizing gameplay balance at levels of play where even reasonably adept players reside.

Now, Sakurai has decided, publicly in the above-mentioned statements, that he's not catering to casuals. He's decided he wishes to design around a compromise between tourney-level players, and the casual playerbase. Players who put enough time and effort into the game will likely, at some point, enter Elite. I finally did after several months of trying to improve.

In my experience Elite represents these tenets of player development:

1. You likely have a dedicated main

2. You likely are aware of tournaments, a few big players, and who's good at playing your main

3. You're likely aware of many of the tactics, tricks, moves, and general gameplay your main can offer

4. You're aware of your strengths and weaknesses, both as a player and as the main you have selected

5. You likely are interested in weeklies or have been to them in the past

6. You're likely aware of tier lists, and are aware of where your main sits on that list

7. You likely have started caring about game balance, as it has presented you with barriers to beating certain people, certain characters, or playing a certain way

This is probably why Sakurai has decided to look into Elite's data specifically, as opposed to all the matches queued online, or as opposed to the representation of various characters in tournaments. People not yet in Elite may not have hit all seven of these points (or most of them), and aren't yet ready to stress over gameplay balance issues to the degree people in it may.

He's balancing around the largest sect of players he can that actually care about balance.

Now, I do have a few concerns with his decision to balance in this manner, but it has nothing to do with the 2.5% vs 1% figure.

1. Online does not play like offline. Online suffers latency delays and game stutters, which, regardless of internet quality, are always present to some degree. This hurts some characters disproportionately more than others, and may skew results.

2. All players are participating in an Elo matchmaking system. One would believe a better statistic to observe in this case would be the average rating of each character, rather than their win rate. A win rate should be 50%, regardless of the relative efficacy of that character. Where you'd expect to see a discrepancy is in the average Elo rating for each character.

3. Balancing around Elite does leave the tournament scene out. This can have a negative impact on the Esports scene when it comes to Smash, as the highest level players are capable of doing things most players, even most Elite players, likely cannot do effectively.