Greenlit: Pokémon's Level Curves

[ddrox13]

Pok?mon's Level Curves

Because I'm literally 20 levels underlevelled and I really don't appreciate it.

Anyway, the plan here is discuss how the levels increase as you progress through Pok?mon games. What makes a good curve is really up to the player, but I figure I'd voice my thoughts on them. Of course, this is a non-TCG ddrox article, so there will be some Real Numbers.

First, I'd construct the actual level curve for each game in the series. This will be done very statistically. As in, I'm making a scatterplot with the routes on one axis and the median level of the Trainer's Pokemon on said route. I would then tell Excel to find the best fit curve, which would serve as the actual level curve. A similar one would be created with the wilds of each game.

Here is a sample chart after I randomized some data:

Still debating how to make the chart look best, but you get the point. Also that data is awful.

EDIT: Should be noted that my current plan is to have all the graphs in a G Sheet or something and link that, but only post in the article ones relevant to whatever discussions I come up with.

After this, I'd run some comparisons discussing some things that come up in my mind often, like...
Why does Johto get a bad rap for it's levels? Is it just Red?
What is up with the wilds on Mt. Cornet?
Are Gym level spikes Good or Bad?
Originals vs 3rd games vs Remakes - which were really harder?

Plus anything else I notice going over the data.

 

[bobandbill]

Neat idea imo. Some scattered questions.

Some thought will be necessary on comparing the level curves of different games as they are of different lengths. GSC to reach the champion is shorter, but then there's all of Kanto. Speaking of, comparing GSC with HGSS would be really interesting (I see this will be a point already including third versions).

Where are you going to get the data from/how?

A note too to be careful how you do the fits (such a thing as overfitting the data, etc). I can also do coding of figures outside of excel if provided the data (and fitting - yay Python), but that's up to you. Google sheets is fine enough.

 

[ddrox13]

Neat idea imo. Some scattered questions.

Some thought will be necessary on comparing the level curves of different games as they are of different lengths. GSC to reach the champion is shorter, but then there's all of Kanto. Speaking of, comparing GSC with HGSS would be really interesting (I see this will be a point already including third versions).

Where are you going to get the data from/how?

A note too to be careful how you do the fits (such a thing as overfitting the data, etc). I can also do coding of figures outside of excel if provided the data (and fitting - yay Python), but that's up to you. Google sheets is fine enough.

Thanks!

The plan was to go legit just count areas. That said, most of the comparisons are either with the same game (wilds vs trainers and stuff) or with other games from the same region (GSC vs HGSS being one big one). BW vs BW2 could be a problem there for obvious reasons, but that's a problem for later.

Plan was originally to just pull the numbers manually from the Serebii pages for each route.

Fitting wise, the sample I came up with I used a polynomial best fit line with a max power of 6, aka the curviest Excel would let me do. This is up for change.

 

[Charlie Brown]

Don't have any other questions but I really like this idea, would be really interesting. There's a lot you're wanting to tackle, so might suggest having a series of articles rather than just the one - though not sure how you can split it up. Just don't want the reader to be overwhelmed with all the graphs and data and analysis in one go.

 

[bobandbill]

Don't have any other questions but I really like this idea, would be really interesting. There's a lot you're wanting to tackle, so might suggest having a series of articles rather than just the one - though not sure how you can split it up. Just don't want the reader to be overwhelmed with all the graphs and data and analysis in one go.

Series, or breaking it up into multiple pages is another alternative.

 

[SirBoglor]

Series, or breaking it up into multiple pages is another alternative.

I'm fine with either one.

The idea is sound enough with some statistics that should be very interesting. Hope that it all goes well.

+1 assuming that you don't ddrox it too much (don't make anecdotes every other sentence plz)