Welcome, Guest. Please Login or Register    
 
smk mk64 mksc mkdd mkds mkwii mk7 mk8
general   mafia   smk   mk64   mksc   mkdd   mkds   mkw   mk7   mk8   |   problems   |   discord   irc
 
  Home Search Members Login Register
 
Pages: 1
Send Topic Print
MK Competion: A Quantitative Analysis (Read 291 times)
Leone
Titan
*****
Offline

Curious.

4390 days karting
København, soon
Gender: male
MK Competion: A Quantitative Analysis
10/04/14 at 15:34:11
 
Hi; I won't waste time with introductory crap, so:

There are a lot of differing opinions as to which MK is 'hardest' to master. I decided to give a go at comparing the PRSR's of each game, to construct a table outlining competitive skill concentrations, by which can be measured the difficulty of a game's chart ladder. I'll add comments to this post explaining extreme differences as you guys ask, or as you make your own comments, but most games' data is pretty easily explainable, so I'll assume y'all already know alot about why this is the way it is, and not waste words adding obvious information.
_________________________________________________
There are a lot of potential uses for this data: For example, comparing PRSR and ARR, with a base ARR system as 'ideal', would allow more balanced standards to be created in the future. (I'm guessing MK64 is closest to being 'ideal', though you guys would know better than I, if there's a game with more solid standards.)

Here's the PRSR data as of 10-1-14.



Numbers on the graph correspond left-to-right with each MK on the spreadsheet.



Edit-- Here are the percentages by game.



The 15th column on this chart corresponds to the percentages of players at each level across all 8 games.



The only flaw I see in this data is in the total comparison, as some games are listed twice or even three times, to take into account every ranking system, which makes the total results weigh more heavily to those titles. If we could all agree which chart in each game is superior, I'd be happy to remove the inferior tables to rebalance the averaged results.  Smiley

It would be waaaaay easier to extract the PRSR data for all 8 games if the combined ranks ran off that system.

More to come maybe.
-L
Back to top
 
« Last Edit: 10/04/14 at 19:20:00 by Leone »  

MKDS #21 WW #3 US
MKRecords Domain Co-Chair
View Profile   IP Logged
Andrew K
King
****
Offline

The Martyr of SMK

worships 4729 kartgods
Jackson, Mississippi
Gender: male
Re: MK Competion: A Quantitative Analysis
Reply #1 - 10/04/14 at 16:31:02
 
Interesting data, definitely does a good job showing which games are the hardest to master aka reach near-WR times on. Also kinda sad to see how the new games are getting progressively easier; just look at the pyramid of zeroes in the bottom right corner of the table. I'd hate to see it continue but I can't imagine Nintendo breaking the trend.

I've always thought that MK64 God times were a little too easy. I suck at the game yet have 2 God times (SC but still), whereas in my best two games (SMK and MKDD) my highest standards are a Titan C (though Legend C non-NBT represents me better) and Titan+ respectively. MKDS and MKDD both have very hard but still obtainable God ranks; each game only has a couple Gods which I think is the best way to have it. SMK and MKSC are both approaching their first overall God, which is a bit too difficult imo. The three latest games are nowhere close to having an overall God, but MKW is the only one that's more than a few years old.

Thanks for screwing up the SMK PAL data with your recruits, Sami. Main reason I made this post Grin
Back to top
 
 

SMK: #60 NTSC and #11 NTSC Non-NBT (USA #15 and #2); 5 non-NBT USA Records (Videos Here)
MK64: #211 (USA #77)
MKDD: #451 Combined (USA #137)

First American to sub 9 seconds on VL2 in SMK
View Profile   IP Logged
AlexPenev
Master Ninja
*****
Offline

I play in 70Hz

punched 7997 kiddies

Re: MK Competion: A Quantitative Analysis
Reply #2 - 10/06/14 at 19:57:09
 
Here's an alternative view of the prsrs:


  • mk7 and mk8 are "easiest" here. This could be either because there's fewer techniques to master or because these games are newest and the top players haven't had enough time to stand out from the pack yet.
  • mk64 (orange) seems 3rd "easiest" initially, but it also had the steepest drop after half-way down the ranking. Perhaps a large portion of dedicated karters near the top of the ranking have played this game for longer periods of time and improved their timesheets over the years. The steep drop off may also be all the players who never tried SSMTs.
  • mkw (dark blue) starts off lowest because it has too many different WR holders. It also has some fairly low values in general because 1/3rd of the tracks have shortcuts.
  • mksc seems quite hard since the drop-off is sharp, and few of the top players have above 0.9. However, it also has the smallest number of players.
  • smk, at least PAL, seems quite challenging considering its curve is quite low and is lower than MKW (which has some tracks with shortcuts where PRSR scores are worse than average)


Some obvious caveats:
1. single exceptionally good players scoring the bulk of WRs can push whole curves down a few notches, so the slopes may be more meaningful than which curves are higher or lower.
2. you also need to consider the number of players on the charts, e.g. a popular game with many casuals is not the same landscape as a niche game where the bulk of players are serious and dedicated.
Back to top
 
 

chown -R me ~you/base*
View Profile   IP Logged
Shock
NinjaShock
*****
Offline



5698 equations solved
North Carolina
Gender: male
Re: MK Competion: A Quantitative Analysis
Reply #3 - 10/06/14 at 20:37:06
 
I enjoyed looking these over. I think MKSC looks unusual mainly due to the number of strats (primarily shroomstrats) spread across across the courses that are hard to pull off but save a ton of time. For example, if you don't use the RR WR strat, you're probably looking at a time that is 8 seconds slower than the WR (which is barely over half a minute long). There's also other difficult, "major" strats on a number of other courses like CCI, SG, BP, etc.

Add this on top of the fact that most players don't have ZZMT and haven't tried the dozens of new strats that have pretty much continually appeared for a while now... and stuff like half a second being knocked off a 5 second WR with no strat change... and the fact that Chris and I have improved the WRs nearly 200 times this year already.... There's a lot of factors contributing to the steep slope of MKSC here.  Smiley
Back to top
 
 
View Profile andylundeen   IP Logged
Leone
Titan
*****
Offline

Curious.

ate 4390 donuts
København, soon
Gender: male
Re: MK Competion: A Quantitative Analysis
Reply #4 - 10/06/14 at 21:24:57
 
Nice graph Alex; is that a function of the database? Because it looks like every player's score is mapped, and I can hardly imagine the work of entering 3.5k scores into a spreadsheet Grin

I'll post some PRSR/ARR comparisons in awhile, in the form of PRSR*|ARR|, where |ARR| ('absolute ARR') is each game's ARR adjusted to a set range of values between 0 and 1 (0 being god, 1 being newbie, and the other major divisions on the graphs being in the form of 'X ARR/Worst Possible ARR').
Back to top
 
 

MKDS #21 WW #3 US
MKRecords Domain Co-Chair
View Profile   IP Logged
Etch
sage
*****
Offline

3.0+1.0=CUTE

6964 time runner-up

Re: MK Competion: A Quantitative Analysis
Reply #5 - 10/07/14 at 13:20:41
 
Here's the delta b/w #1 and #10 for prsr....



Like Shock said, hitting the wr strat is critical for mksc and they are not easy. Tongue

I'd like to think with more popular games the top 10 is representative of hard core competition.  Of course, that doesn't mean reaching WR's are easy but there is less reliance on strats that require luck for lack of a better term.
Back to top
 
 

Karters cannot create anything out of nothingness. Karters cannot accomplish anything without holding onto something. After all, karters pretend to be gods.
View Profile   IP Logged
AlexPenev
Master Ninja
*****
Offline

I play in 70Hz

ate 7997 donuts

Re: MK Competion: A Quantitative Analysis
Reply #6 - 10/07/14 at 15:14:02
 
^ perhaps difficulty is not uniform and we can look at it from different viewpoints in each ranking. You're checking the improvement needed to go from 10th to 1st but we can pretend to be in any rank and ask "what would it take to halve my rank". So to move from last place to middle place (x=1 to 0.5) might be hardest in MK64 and easiest in MK8, since they have the steepest and gentlest slopes in that section. To go from 50% to 25% might be harder in MKSC and easier in MK7, perhaps due to some techniques that need to be learned? Likewise, for someone at 25% up the ladder to improve to top-10% seems hardest in MKSC and SMK-PAL.

But we gotta be a bit careful when comparing PRSR scores directly because some games have shorter tracks. Making half a second of errors in a five minute race is not as damaging to PRSR as making half a second of errors in a 50 second race.
Back to top
 
 

chown -R me ~you/base*
View Profile   IP Logged
Vinnie927
King
****
Offline

#1 lurker

4476 days karting
Canada
Gender: male
Re: MK Competion: A Quantitative Analysis
Reply #7 - 10/07/14 at 15:15:15
 
Maybe total times/AF could be used for this as well
Back to top
 
 
View Profile   IP Logged
Leone
Titan
*****
Offline

Curious.

punched 4390 kiddies
København, soon
Gender: male
Re: MK Competion: A Quantitative Analysis
Reply #8 - 10/07/14 at 16:01:45
 
Vinnie927 wrote on 10/07/14 at 15:15:15:
Maybe total times/AF could be used for this as well


Total times weigh too heavily to long tracks; really not optimal

And AF curves heavily in games with a small no. of players. I doubt Andy'd have 1.1 AF in MKSC if there were as many players as on '64 or 'DD, for instance.
Back to top
 
 

MKDS #21 WW #3 US
MKRecords Domain Co-Chair
View Profile   IP Logged
Etch
sage
*****
Offline

3.0+1.0=CUTE

6964 time runner-up

Re: MK Competion: A Quantitative Analysis
Reply #9 - 10/07/14 at 16:50:56
 
Leone wrote on 10/07/14 at 16:01:45:
I doubt Andy'd have 1.1 AF in MKSC if there were as many players as on '64 or 'DD, for instance.


Unlike 64 and DD, there is a shortcut aspect to non-sc here.  There are only a handful of players that have landed WR strats on specific tracks.  Sure, we may have the fewest overall competitors but I assure you it's top of the crop!
Back to top
 
 

Karters cannot create anything out of nothingness. Karters cannot accomplish anything without holding onto something. After all, karters pretend to be gods.
View Profile   IP Logged
Blackyboi
King
****
Offline



fell 4679 times for marty
Austria
Gender: male
Re: MK Competion: A Quantitative Analysis
Reply #10 - 06/20/15 at 16:54:13
 
Hardest MK to optimize all around? MKW with sc since there are simply so many different strats which also require a lot of practice, hardest MK overall, I mean just look at MKSC sc numbers, but then again I think as already mentioned it also depends on the number of entries in a certain game.

Overall I think it's really hard to compare the games tho, since all require a lot of work to be optimized just to the standards we have as Top times now, the probably "easiest" MK is MK7 since there are no really hard techniques in terms of mechanics. Also I think it can't be said generally since for some players one game might be/feel easier and more fun than for others which is why they mostly stick to it.
Back to top
 
 

ALAKTORN wrote on 04/27/13 at 08:29:36:
it’s funny to see that Walter is always right in the end

View Profile   IP Logged
Pages: 1
Send Topic Print