Smash 4 Match Data Collection

[BestTeaMaker]

So, there have been a lot of debate about 2v3 stocks, the time limit, if customs moves help, how good characters are, how good certain players are, etc etc...

So, instead of merely saying words and such, why not actually collect match data?

I present the Smash 4 Match Data Collection table!

This has nearly anything I can think of in terms of collecting character data with some feedback according to what is input. Eventually, after enough data is collected, I'll create an infographic so that people can have a nice, visual presentation.

I'll be constantly updating this and possibly will add future data points concerning matches. If you're interested in helping out to collect data, give me a pm and I can add you. Hopefully, someone will use this information and put it to good use.

 

[BestTeaMaker]

Progress Report 1:

Data input - Dynamitti IV, KiT 15

So, I think I spent way too much time formatting this table, but goddamn do I love how it looks, especially with the number of janky workarounds I had to work with.

In terms of features of this table, there are quite a lot. They are as follow:

Match Information

• Source link
• Event name
• Players
• Version
• Mode of play
• Region
• Date
• Round
• Customs used or not
• stock count
• time

Game

• Winner/Loser
• Character choice
• Time end
• Game length

Match Summary

• Winner
• Score
• Match length
• Games played
• Time per game

Now, time to go and update the character images, then I can continue to collect information. If you're interested in joining me on this venture, shoot me a PM.

 

[BestTeaMaker]

Progress Report 2

So, I just finished uploading the matches from VGBC for KiT 15. Here's the following stats for KiT:

Total Matches - 40
Total Games - 99
Average Time per Game - 4m 12s
Standard Deviation - 1m 21s


Here are the following stats for all the 3 stock data I have so far:

Total Matches - 55
Total Games - 143
Average Time per Game - 4m 20s
Standard Deviation - 1m 21s

I also added the following data point:

• Stages

If you have any suggestions, or wish to help me in this endeavor, let me know.

 

[Dr. Tuen]

This would require more careful combing of the information you have so far, but trying to assess time per stock would be interesting. For a two stock match, that'd mean recording 2-3 match times (stock losses) per match.

I'm also personally curious about the frequency of comebacks. When does the person who loses the first stock WIN the match? I think that's an interesting part of the 2-3 stock argument, since some would argue that 3 stocks allows comebacks and read time for both players.

 

[BestTeaMaker]

This would require more careful combing of the information you have so far, but trying to assess time per stock would be interesting. For a two stock match, that'd mean recording 2-3 match times (stock losses) per match.

I'm also personally curious about the frequency of comebacks. When does the person who loses the first stock WIN the match? I think that's an interesting part of the 2-3 stock argument, since some would argue that 3 stocks allows comebacks and read time for both players.

Boy, more work!

But yeah, I was thinking about looking over how stock progression goes for each match. I mean, hell, since I'm going back in to add stages used in each set, might as well see the stock count.

 

[ZFallen]

Great job man, this is what we've been lacking with Smash 4! If people wish to discuss rule changes and such, having quality data will help us make quality decisions!

 

[Dr. Tuen]

Boy, more work!

But yeah, I was thinking about looking over how stock progression goes for each match. I mean, hell, since I'm going back in to add stages used in each set, might as well see the stock count.

Yeah, if you do stock time/count, I'll be real interested to apply statistical analysis of sorts to it. Though I can't say I have an abundance of time on my hands, given I'm in the home stretch of grad school. But if I get enough time, and this data is here, I would love to check it out!

 

[Thundering TNT]

Keep up the good work. This could ge interesting, and if anyone wants to affect change, having good, concrete data to back them up will really help.

 

[BestTeaMaker]

So, I tried tackling the task of recording stock data for each game. 2 important questions to ask with this:

• How long does a stock last for each character?
• Does the person who loses the first stock lose the game?

Currently planning on ways to do this. Hopefully something will get through to my head. It's hard to do this alone.

This comes to my next statement. I can't really do this alone. It's gonna take a LONG time for me to search for tournaments and input the data myself. So, I'm asking a general plea for help in filling this out. If at any time you're watching a tournament through a stream, VOD, Youtube, etc...just simply take this table and fill it out as you go about watching the match. I've made the sheet as easy as possible to work with, locking certain aspects so that data doesn't go missing.

So yeah, if you want some hard data for the competitive meta of this game, please help me collect it.

 

[BestTeaMaker]

Progress Report 3

Woo! I made a table that can answer the two questions I posted earlier. Making that table took all night to figure out, but here it is!

 

[Overswarm]

Interesting. I did something similar with Brawl in the past.

 

[BestTeaMaker]

Progress Report 4

Been going through KiT15 against slowly to get match data. I swear, this tournament alone has made me hate Smashville more and more. STOP COUNTERPICKING THERE. IT'S NOT GONNA HELP YOU. Apparently March 27th is the deadline for me if I want to make any headways concerning EVO, so I gotta get into gear with data collection.

In the meantime, because I'm a crazy man, I added the following data points:

• Stock Data
• Matchup Data

The Matchup data sheet is purely based on tournament results. There are some weird missing numbers because math doesn't like a record of 2-0, but hopefully as I slowly fill this up, those spaces'll get populated with more numbers.

 

[Overswarm]

Good stuff. I don't have much time for data collection, but let me know if you need any help with analysis Tea.