Pokémon Can we talk about learning curves?

[Lojh]

how many different types of curves are there?
Are there different curves for different generations?
Is it possible to assess my skill 4, 5, 6 years from now?
How easy are some generations to pick up?
how easy are some generations to Master?
Can we graph learning curves of past players to create an estimate for a new player?
Why do some people plateau whilst some people continue to develop?
Is age a factor in learning curves?



Mathematicians, get number crunching.

 

[Sevi 7]

I actually was just discussing this with someone about this last night. So, I'll talk a little about it.
A couple things to note though. You said gen specifically, but I'm going to talk about tiers too. Even between an OU and UU tier there can be huge differences. Also, I'm going to mostly use Gen 1 OU as an example, because that's generally considered the "main tier" on here.

how many different types of curves are there?
Are there different curves for different generations?

Short answer: Yes.
Long answer: Just from the perspective of what one needs to know to be decent at the tier, every tier has different things that are more important to learn than others, even if the mechanics are the same. If we quickly compare Gen 1 OU and Gen 4 OU, for example, being able to predict an opponent's team/core is vastly more important in Gen 4 OU.
When looking at "mastering" the tier, the differences become so much different that there's not even a comparison. Gen 1 OU has a learning curve of playing around Reflect Chansey, which obvious doesn't exist in anywhere else.

Is it possible to assess my skill 4, 5, 6 years from now?

Short answer: No.
Long answer: People grow at different rates, hit different plateaus and get through said plateaus at different speeds (some never get over certain plateaus as well). Sometimes it has nothing to do with how much plays or trains, but it has to do with other outside factors, even genes/personality. People are too varied to make an accurate guess half a decade down the line. Especially if the goal is to make a one size fits most kind of guess.

How easy are some generations to pick up?

Short answer: Most generations are going to be the same, give or take. There are definitely people who will fight me on that though.
Long answer: I would say that all gens have a lot to learn and are deceptivly deep. As for what is easier, that all depends on where you come from. I started playing long before team preview, so that's not something I have to worry about. I already understand the lead meta, etc. However, for a lot of players who came in durring XY, SM or even BW, this can be another hurdle (or learning curve, since we're on the subject), making these tiers harder to pick up for them.
At the same time, I don't play a lot with Fairy-types and so it's not something I really remember well, and other types like Poison have a much different utility in my mind. This is fine when I play ORAS NU casually, but it's definitely something that keeps me from playing at the next level. So, even an old gener vs new gener kind of mentality doesn't work.

how easy are some generations to Master?

I think you have to define master, because everyone will probably have a different perspective on how that word should be defined. Thus meaning that you'll never be able to get an answer as people are constantly talking past each other.

Can we graph learning curves of past players to create an estimate for a new player?

Short Anwer: Maybe, but with a lot of work and imperical evidence, which is effort better spent elsewhere.
Long answer: Scientifically speaking, in order to adjust for different people's different backgrounds, interests, personalities, etc. we'd need hundreds of players to learn in an controlled enviroment, and chart their progress. We'd also have to figure out how we wish to define different progress points and what would be important information to graph in the first place. Even if we used peer-reviewed data involving mastering a skill, etc. we'd still need to do ground work on figuring out how that would best apply to pokemon. If you really want an accurate chart, that would be years of work with no compensation. I guess you could go to your local university and see if a Pokemon loving Ph.D student would want to do their disertation on something like this. But I don't a review committee would pass them, and being a Ph.D student is already hard enough.

Now, if you really want your chart and don't care much about accuracy. I guess you could make random guesses, based on anecdotal evidence. But then, what was the point in making the chart in the first place? It's not going to be accurate and isn't actually useful.

Why do some people plateau whilst some people continue to develop?

Short Anwer: Everyone plateaus. How, when and why all depends on a long list of factors ranging from genes to interest to previous experience.
Long answer: Everyone is going to plateau at all the things they do. If you haven't plateaued at something yet, then you haven't done the best you can yet. Why? Well, there are so many reasons and factors involving why one may plateau at a certain point, that you have to look at it from a case by case basis. You may be able to make list of generalities, but they'd have to be vagued, and person would have to be honest with themselves on a persoanl level to figure out which ones kinda apply to them. That's going to be the equivilent of Buzzfeed and Facebook quizes though.

Is age a factor in learning curves?

Short answer: Yes. However, it may not be how you think, and that doesn't mean you can't get over it.
Long Answer: The way age affects one's ability to learn has to do with how developed the brain is (the brain stops developing around the age of 24-26, but the brain has different stages at which somethings are more optimal to learn at and is generally speaking the younger the better) and how much ability the brain has lost -like when your great grandmother can't figure out how they get electricity on planes. My psych background is with linguistics, so I'm not an expert on when exactly which periods are best to learn a strategy game. However, there is probably information out there. Just use Google Scholar (and if you need help reading a scientific paper I could quickly explain it).

However, sometimes thinking that age is stopping you from learning something can be more detrimental than the age itself. Your preception and your brain's reality are sometimes the same thing, even if it differes from actual reality. If you think you can't learn something, your brain may perceive that as the reality of the situation and you won't learn it. Why should your brain waste precious resources trying something that it's going to fail at?

Mathematicians, get number crunching.

Just fyi, this is absoutely something for psychologists, not mathematicians. Yes, it involves stats, but every psychology undergrad is going to take multiple classes involving, learning and using stats. There's not a psychologist, worth their salt, that can't take this information and chart it, and everything outside of actually making the graph is psychology.

EDIT: I'm sure there are mistakes, but I don't really want to proofread this. Hope that's ok, and sorry if it isn't.

 

[Lilyhollow]

The way I think of a learning curve is that at the beginning of the curve, you tend to be learning many things--but things which tend to be relatively cognitively shallow. As you ascend the curve, there's less to learn, but the stuff remaining gets deeper and more difficult to improve at.

Some shallow things about Pokemon that people have to learn initially are the type chart, the various stats of each Pokemon, what the moves do, general mechanics, etc. How much of this is present is highly dependent on the format. So if you're playing Pokemon for the first time and you're starting with gen 1, it's pretty easy to learn this stuff because you can just look up resources that will flat out tell you that only 20 of the 150 Pokemon are relevant. There are fewer moves, simpler moves and mechanics, and a smaller type chart. You can really really quickly get to the point where you're learning the deeper stuff.

Sometimes there is a lot more 'lower-level' stuff to learn, and it can bog you down from learning the higher-order skills before you even really get a chance. I think this is actually one major reason (of many) that some people plateau early. For example, in Sun/Moon or whatever, there are going to be way more than 20 viable Pokemon I need to learn about, way more moves to keep track of, more typings, sometimes frequent bans/unbans that keep the metagame even more in flux, and so on. There's tons to learn, but most of it is really just the lower-level stuff, distracting from one's ability to improve higher-level skills.

Physical sports (or real-time games) are a good comparison. There's not really that much, cognitively, that goes into being athletic per se. Like I could start running miles every day and as long as I avoid injury, eventually I'm going to be pretty fucking good at it. I of course have no desire to do that whatsoever! But if I want to play a fun game like basketball, and learn a bunch of basketball skills, this will actually end up being a really big problem for me; I may want to improve some of my slightly higher-level skills like, I don't know, 'learning how to get into the right position,' but how am I ever going to do that if I'm physically too slow to ever effectively do that? The reality is that I need to be running laps! But some people may not realize this. Back when I consistently played Pokemon (Gen 3), I never really did. I wanted to try learning the higher-level stuff right away, but there was a lot of basic stuff I was just too lazy to put any time into, so I plateaued very early.

Regarding 'mastery,' we're kind of on a spectrum here so I don't feel that it's entirely useful to identify some arbitrary point at which someone achieves a 'mastery state.' Instead I conceive of mastery as a continuous process of improvement that does not end until one simply decides that further improvement is no longer personally satisfying. Which can be based on a whole bunch of different things.

Love the thread idea btw!

 

[Lojh]

Ah i realize master is a relative term. How about i take a page out of Callous’s vocabulary and classify master as

“SPL-Level”?

 

[Lilyhollow]

So at that point you're talking about relative skill level. Things get really muddy there because so much is determined by various features of the playerbase, such as its size. Becoming a top 100 pokemon player is gigantically easier than becoming a top 100 chess player, but that really has little or nothing to do with how much there is to potentially learn about the games.

So my suspicion would be that classifying any SPL-level player as a 'master' would result in it being harder to become a 'master' of S/M than RBY OU. But if you multiplied RBY OU's playerbase by 10000, all of a sudden it's like basically impossible to 'master' the game. Which doesn't really seem right.

My definition of a 'master' would likely be something along the lines of 'someone who has achieved a high level of competence at all the major skills that have been codified and deemed requisite for having a full understanding of the game.'

The limitations of this definition would be:

• Requires such skills to be codified by the community
• There's a lot of subjectivity because 'at what point do we deem someone to be highly competent at something'
• Still is based on some level on community attributes such as 'how well does the community even understand the game'

But the good thing about the definition is that it avoids turning 'mastery' into a totally relative goal. Like someone in 1910 who was considered a 'chess master' would get totally destroyed in today's game. It's good to give them some recognition of their ability which is relative to their own time period and playerbase, but it would be quite disingenuous to refer to them as having 'mastered' the game of chess as we know it today. It would be more useful, I think, to decide that someone has only mastered the game of chess as we know it once they have demonstrated knowledge and performance at the skills which we as a community understand to be important for mastery of the game.

So by this definition, 'mastery' would occur somewhere around the rightmost part of this learning curve. Where exactly it would occur would still have some element of subjectivity, which means that it's still hard to say if one format is 'easier to master' than another. After all, a particular community could simply raise or lower the standards of what constitutes mastery.

But the advantage of this is that, if we define you as a master of Gen 2 OU in this way, it would actually imply a lot of specific things about you as a person. For example, one might imagine that you were required to demonstrate certain highly-developed battling skills, while maybe the game's lack of variety means you can be just 'kind of okay' at certain teambuilding skills and still have mastered the game. Meanwhile, in S/M, perhaps it's actually necessary to reach a really high level of attainment in teambuilding skills to be considered a master of the format. So we could then predict that a Gen 2 OU 'master' who tries their hand at S/M OU would probably pick up battling skills quite easily, but may or may not struggle at teambuilding. Based on this, we could maybe make a prediction of what the average Gen 2 master's learning curve would look like for S/M.

Umm, that's all i've got lol. I don't know if any of this made sense!

edit: oh one last thing. Another nice thing about defining mastery in this way is that we can apply it more usefully to new or unfamiliar games. For example, it probably makes no sense to refer to anyone as having 'mastered' Gen 3 3U right now (i'm pretty sure it doesn't exist). We could invent it and hold a tournament for it today, and just declare whoever wins the first tournament to be the first 'master' of the format, but of course we all know that's bullshit, specifically because we know what mastery will generally look like one day due to our experiences in other formats. So this is another regard in which looking at relative skill or raw elo ratings is clearly not a useful basis for declaring mastery.