Logo

Tournament Scoring

註冊賬號 登錄 Back To Forums

Post #1 · Posted at 2012-09-07 07:08:20am 12.8 years ago

Offline Dr.D
Dr.D Avatar Member
1,451 帖子
United States
Reg. 2008-03-01


Last updated: 2012-09-07 09:33am
In this thread post what you feel is the best method to determine the winner of a song. I also wouldn't mind if you post your idea for how to decide the winner of a match. (like 2/3, round robin, etc.)

I'm mainly curious about how you would count this on new DDR games like SN2 and now with the new scoring system. I already think there's a strong consensus with older games.

My view in a nutshell

In the west, we have never used machine score to determine winners from Extreme and lower. It simply wouldn't have been fair or made sense. It seems now with the new scoring system people are willing to give up calculating scores in favor of just looking at the machine score. I do not find that fair for a few reasons.
Before I get into that, I think it's important when having this discussion to state what you are looking for when your having a tournament. As in, do you care about who has the best accuracy, or do you care more about who gets the less amount of greats? If the latter is what you care about, this thread wont be of much use to you. I've always thought the former is a far better way to determine the winner.

A good supporting point of that would be to look at any IR ranking course. Konami knows not to use machine score during an IR course. It should be quite obvious why as well. If they were to use MS (machine score) during a course, it would always come down to who had less greats and it wouldn't make much difference how many marvelous or perfects you had. If you still don't see the point and think it would be fairer to use MS for an IR course then I am afraid there probably wouldn't be any way to convince you otherwise.

For the record, last year at KAC2011, the matches were not decided by MS, Konami used their DP system instead which I was pleased to see.

Does that mean I think the DP system is the best? Not necessarily . I'd be interested if some of you can help come up with a better idea to how to score tournaments in this new age with marvelous timing. So go ahead, express your opinions.
Add my DDR 2013 Rival code-1106-3749
Visit my YouTube channel DDR 2013 videos!
DrDvids

http://zenius-i-vanisher.com/ddrsig/2647.png?t=1235917793

Post #2 · Posted at 2012-09-07 09:27:16am 12.8 years ago

Offline Zowayix
Zowayix Avatar Member
1,149 帖子
Not Set
Reg. 2009-09-19

I thought IR courses were decided by machine score. However, courses do use a significantly different scoring system than normal mode, in which there is just a large a difference between Perfects and Marvelouses as there is between Greats and Perfects, making it very similar to the DP system and once again fair.

Post #3 · Posted at 2012-09-07 03:09:43pm 12.8 years ago

Offline seishinbyou
seishinbyou Avatar Member
617 帖子
Japan
Reg. 2010-07-08

Game Center Nickname: aijbot
"Wow, 4 types of corn!"
Yeah, nonstop scoring is slightly different, though for IR purposes everyone is on the same plane. Also with nonstop scoring in the X series, note that steps in later songs in the course are weighted more heavily than those songs in the beginning of the course.
http://aaronin.jp/shashin/konmainewbanner.jpg

Post #4 · Posted at 2012-09-07 03:54:23pm 12.8 years ago

Offline Parafox
Parafox Avatar Member
111 帖子
Not Set
Reg. 2011-09-18

I think deciding by machine score is the best way to go. The reason being is that the "Marvelous" rating is just a tie breaker for PFC's. For instance if you played DDR X2 songs on a DDR Extreme cabinet, the Marvelous wouldn't show and obviously whoever got the closest to an AAA would win. In DDR SN2 - X3, I really enjoyed that they kept the whole "PFC is still better than a black flag" type thing. For instance you can get more Marvelous ratings than the other person, but if you got a great that would still reduce you down to an AA in SN and below. Overall, I like the machine score because I think it still stays true to the original arcade scoring system while being easy to recognize what percentage you got.

Post #5 · Posted at 2012-09-07 05:44:57pm 12.8 years ago

Offline Dr.D
Dr.D Avatar Member
1,451 帖子
United States
Reg. 2008-03-01

So would you also suggest IR courses to be determined using machine score?
Add my DDR 2013 Rival code-1106-3749
Visit my YouTube channel DDR 2013 videos!
DrDvids

http://zenius-i-vanisher.com/ddrsig/2647.png?t=1235917793

Post #6 · Posted at 2012-09-08 05:12:35am 12.8 years ago

Offline seishinbyou
seishinbyou Avatar Member
617 帖子
Japan
Reg. 2010-07-08

Game Center Nickname: aijbot
"Wow, 4 types of corn!"
Quote
I really enjoyed that they kept the whole "PFC is still better than a black flag" type thing.

I think that is why they kept the colored circles to indicate FC/PFC/MFC actually, because it is still possible for things like a black flag to have a higher score than a PFC with enough perfects.

On a per-song basis, I'm actually impartial to how tournament scoring is done on later mixes if they use machine score or any kind of dance point system.

If we were talking about mixes like Extreme, then I definitely think DP is more fair, and for mixes like 5th, definitely use DP since mixes like that still value full combos too much over score. (You got a lower score regardless if every step but one (Good or lower) was a Perfect because of the bonus system)
http://aaronin.jp/shashin/konmainewbanner.jpg

Post #7 · Posted at 2012-09-08 06:09:05am 12.8 years ago

Offline Parafox
Parafox Avatar Member
111 帖子
Not Set
Reg. 2011-09-18

Quote: Dr.D
So would you also suggest IR courses to be determined using machine score?

Well I consider myself a very casual DDR player (only a few arcade PFC's), so normally I just go by the machine score. I think it just fits me so that was just my opinion. On the topic of IR, maybe not. IR is a pretty competitive in my opinion, so I guess a different grading system would be required. For instance, since I'm a casual player, I only limit myself to Perfects and whatnot. Since more competitive players enjoy getting a MA, I assume it would only be necessary to count them as more than perfects.

And yeah, you can get a black flag and still get better than a PFC, but you'd have to have a pretty bad PA in my opinion, or that's just my opinion. Tongue

Post #8 · Posted at 2012-09-08 06:36:57am 12.8 years ago

Offline TsukiyoX
TsukiyoX Avatar Member
411 帖子
United States
Reg. 2009-02-15

Machine Score. They (in my opinion) perfected the machine score based on the value of four components (Marvelous, Perfects, Greats and OKs). I used it for my tournament and it turned out fine.

Post #9 · Posted at 2012-09-08 07:27:52am 12.8 years ago

Offline Dr.D
Dr.D Avatar Member
1,451 帖子
United States
Reg. 2008-03-01


Last updated: 2012-09-08 07:40am
As yet, I think we are considering easy cases. Such as, players are evenly matched on the songs and will usually be decided by a great or few. Things get more complicated with new situations. The following are unusual yet plausible cases they may pop up. Please answer if you would make any exceptions in the following cases.

1. It's already been mentioned, but, in extreme cases where a player one has an insanely low perfect count with 1 great and player 2 manages a PFC with many perfects. Actually this one isn't as uncommon as you may think. I often play with my friend DFGH who is quite good though doesn't MA as much as I do. Once in a while I'll botch my PFC with a low perfect count and he'll go on to pfc it.

Before the next one, I want to mention one thing. ANYTHING LESS THAN A MARVELOUS IS A MISS STEP. Well, I mean, if you really want to be honest here. Anything less than marvelous is in other words a mistake. Sure greats are a worse "mistake" than perfects but nonetheless, still a mistake.

So in case one, using MS, the person with the PFC would win even though they have far more mistakes.

2. Goods? Misses? No one ever brings this one up. What if the 2 players tie with a machine score of 996,210. Both players have the same perfect, great, and marvelous count. Player 1 got a legit good where player 2 got a legit miss. According to the machine, the score is tied. But ask yourself, is that really fair? Player 1 actually made an attempt to hit an arrow where player 2 completely missed it without making the necessary attempt. (to be fair, even the DP system can't handle this which is one reason I said before it could use some improvement)

I have a couple more cases but I'll let these ones be discussed first.
Again I'll say, for normal play I wouldn't want to use anything but MS scoring. I would absolutely hate any kind of DP system to display in game mode (unless its some kind of toggle option). The only thing I argue is its not fair in tournament settings cause it will usually hurt a player.
Add my DDR 2013 Rival code-1106-3749
Visit my YouTube channel DDR 2013 videos!
DrDvids

http://zenius-i-vanisher.com/ddrsig/2647.png?t=1235917793

Post #10 · Posted at 2012-09-08 09:41:20am 12.8 years ago

Offline Silverhawke
Silverhawke Avatar Member+
4,606 帖子
Indonesia
Reg. 2009-01-27

3DS Friend Code: 3496-9710-9426
"highwind fluffdragon"
Quote: Dr.D
2. Goods? Misses? No one ever brings this one up. What if the 2 players tie with a machine score of 996,210. Both players have the same perfect, great, and marvelous count. Player 1 got a legit good where player 2 got a legit miss. According to the machine, the score is tied. But ask yourself, is that really fair? Player 1 actually made an attempt to hit an arrow where player 2 completely missed it without making the necessary attempt. (to be fair, even the DP system can't handle this which is one reason I said before it could use some improvement)

That is a very good point. However IMO an attempt to fix this would make the DP system too complicated (unless you want to include negative points again).
my homepage → silverhawke.xyz

Post #11 · Posted at 2012-09-08 12:15:29pm 12.8 years ago

Offline seishinbyou
seishinbyou Avatar Member
617 帖子
Japan
Reg. 2010-07-08

Game Center Nickname: aijbot
"Wow, 4 types of corn!"
Quote: Dr.D
Before the next one, I want to mention one thing. ANYTHING LESS THAN A MARVELOUS IS A MISS STEP. Well, I mean, if you really want to be honest here. Anything less than marvelous is in other words a mistake. Sure greats are a worse "mistake" than perfects but nonetheless, still a mistake.

So in case one, using MS, the person with the PFC would win even though they have far more mistakes.

In terms of raw "number of mistakes", then sure, but a Perfect is far less of a mistake than a Great. So far less (?) in fact that there is only a 10 point difference between the two as opposed to more than half the value of the step difference.

It's up to you if you want to go into degrees of mistakes or just draw a hard line and say "Anything other than a Marvelous is wrong"
http://aaronin.jp/shashin/konmainewbanner.jpg

Post #12 · Posted at 2012-09-08 12:23:39pm 12.8 years ago

Offline Silverhawke
Silverhawke Avatar Member+
4,606 帖子
Indonesia
Reg. 2009-01-27

3DS Friend Code: 3496-9710-9426
"highwind fluffdragon"
That would be like making a percentage judgment...

So there would be base score, and the score of each note would be BaseScore*(100% - [how offset you are from MARVELOUS]/[whole timing range window]) in which the MARVELOUS would be counted as 0ms, and holds would be calculated by BaseScore*([percentage of hold held down]), so if you let go of the hold arrow mid-way you'll get 50% of the base score.

Instead of showing judgments, it shows like 100%, 75% or something like that.

Base score is still equal to 1 000 000/(total notes+total OKs), in which getting 1 000 000 = perfect timing (0ms on all arrow) and holds held all the way.




What the hell am I writing?
my homepage → silverhawke.xyz

Post #13 · Posted at 2012-09-09 10:56:34pm 12.8 years ago

Offline Dr.D
Dr.D Avatar Member
1,451 帖子
United States
Reg. 2008-03-01

This is getting close to what I feel needs to be changed to make the DP system better.
I don't have the timing windows for DDR atm so if someone can post that here I'd appreciate it.

Pretty much I think the points you get for each step should reflect the difference in the timing window.

Example:
Let Marvelous equal the max step value for a song. to make things easiest, let's pretend on said song 1 marvelous equals 1,000 points.

Now, if the timing window for getting a perfect is twice as wide as getting a marvelous, then I think the value of a perfect should be half to reflect it. (I think this is how course scoring works)
Assuming each judgement lower is halved, it should be this
1,000-Marvelous
500-Perfect
250-Great
125-Good (this is where we can fix the problem in case 2 that I mentioned)
0-Miss

Obviously a miss should count as 0 since no step was made remotely close to the step if any. I do feel that goods shouldn't be null though.

Keep in mind I would only want this in a tournament setting where I think the most accuracy matters more and not, "oops, I got a great on an easy song and my opponent will for sure PFC it so I have no more chance."

Anyways, am I just crazy or is any of this making sense?
Add my DDR 2013 Rival code-1106-3749
Visit my YouTube channel DDR 2013 videos!
DrDvids

http://zenius-i-vanisher.com/ddrsig/2647.png?t=1235917793

Post #14 · Posted at 2012-09-09 10:57:50pm 12.8 years ago

Offline Silverhawke
Silverhawke Avatar Member+
4,606 帖子
Indonesia
Reg. 2009-01-27

3DS Friend Code: 3496-9710-9426
"highwind fluffdragon"
That actually makes sense.
my homepage → silverhawke.xyz

Post #15 · Posted at 2012-09-09 11:21:47pm 12.8 years ago

Offline Zowayix
Zowayix Avatar Member
1,149 帖子
Not Set
Reg. 2009-09-19

Isn't that almost identical to how courses have scored it? Yes, it does make perfect sense.

Also, I'm curious about how you say that IR courses are not scored by machine score. In that case, what are they scored by? I thought that the different scoring system in courses that puts a huge emphasis on Marvelous attacking would make machine score fair (compared to the regular scoring system outside of course mode).

Post #16 · Posted at 2012-09-10 03:27:14am 12.8 years ago

Offline Hlavco
Hlavco Avatar Member
113 帖子
United States
Reg. 2009-06-02

Quote: Dr.D
ANYTHING LESS THAN A MARVELOUS IS A MISS STEP. Well, I mean, if you really want to be honest here. Anything less than marvelous is in other words a mistake.

Maybe it's because I don't play with super-high-level people, but I've never looked at it this way. In my mind, Perfect is, well, perfect, and Marvellous is just an even more specific level to be used in the event of a tie. If one person gets the Full Perfect Combo, and the other person gets one Great and the rest Marvellous, then for me the guy with the Full Perfect Combo wins because Perfects are what I'm going for.

But like I said, I play at a low enough level that having Marvellous isn't really necessary anyway.

Post #17 · Posted at 2012-09-10 04:04:53am 12.8 years ago

Offline seishinbyou
seishinbyou Avatar Member
617 帖子
Japan
Reg. 2010-07-08

Game Center Nickname: aijbot
"Wow, 4 types of corn!"
Quote: Zowayix
Isn't that almost identical to how courses have scored it? Yes, it does make perfect sense.

Also, I'm curious about how you say that IR courses are not scored by machine score. In that case, what are they scored by? I thought that the different scoring system in courses that puts a huge emphasis on Marvelous attacking would make machine score fair (compared to the regular scoring system outside of course mode).

I think he just means that IR courses (nonstop/oni) are scored differently than as if they were 4,5,etc. songs played individually back to back. For later mixes:

Quote
The maximum possible score for a nonstop/oni course is the same as for an individual song (1,000,000) and each step score is calculated as follows:

Step Score = 1,000,000 / (Total "dance points" for all stages)

where:

Stage Dance Point Total =
STAGE 1,2 : ( Total number of steps, freeze and shock arrows ) X 3
STAGE 3,4 : ( Total number of steps, freeze and shock arrows ) X 4
STAGE 5,6 : ( Total number of steps, freeze and shock arrows ) X 5
and so on ...

So at the end of each stage your score will be:

Final Score = Step Score * (# of Marvellous and OK steps) X 3
(X4 on stages 3,4 , X5 on stages 5,6 and so on)

+ Step Score * (# of Perfect steps) X 2
(X3 on stages 3,4 , X4 on stages 5,6 and so on)

+ Step Score * (# of Great steps) X 1
(X2 on stages 3,4 , X3 on stages 5,6 and so on)

Note that scores are not rounded to the nearest 10 points as with normal play. Scores are displayed as is except all decimal places are truncated when they are displayed to the screen (the floating point value is always used for internal calculation purposes)
Let's use this to derive a few sample step scores. Take a 5 song course with the following number of steps:

Stage 1 : 250 steps
Stage 2 : 350 steps
Stage 3 : 400 steps
Stage 4 : 350 steps
Stage 5 : 300 steps

So the base Step Score would be:

Step Score = 1,000,000 / (Total "dance points" for all stages)
Step Score = 1,000,000 / ((250*3) + (350*3) + (400*4) + (350*4) + (300*5))
Step Score = 1,000,000 / ((750) + (1050) + (1600) + (1400) + (1500))
Step Score = 1,000,000 / 6,300
Step Score = 158.73015... (will be truncated when displayed to screen)

Now lets use this to calculate the step score for various judgments on different stages to illustrate:

Stage 1 : Marvellous/OK step = Step Score * 3 = 476.19047...
Stage 1 : Perfect step = Step Score * 2 = 317.46031...
Stage 1 : Great step = Step Score * 1 = 158.73105...

Stage 3 : Marvellous/OK step = Step Score * 4 = 634.92063...
Stage 3 : Perfect step = Step Score * 3 = 476.19047...
Stage 3 : Great step = Step Score * 2 = 317.46031...

Stage 5 : Marvellous/OK step = Step Score * 5 = 793.65079...
Stage 5 : Perfect step = Step Score * 4 = 634.92063...
Stage 5 : Great step = Step Score * 3 = 476.19047...

http://aaronin.jp/shashin/konmainewbanner.jpg

Post #18 · Posted at 2012-09-14 09:47:01am 12.8 years ago

Offline Dr.D
Dr.D Avatar Member
1,451 帖子
United States
Reg. 2008-03-01

A good middle ground I thought of could be something like this.

Marvelous is of course the max score for a step. The max score is worth 1 point (1DP). Then the other could be this.
1=Marvelous
.75=Perfect
.50=Great
.25=Good

Not exactly what I wanted but I think this could be a good compromise for tournament play.
Add my DDR 2013 Rival code-1106-3749
Visit my YouTube channel DDR 2013 videos!
DrDvids

http://zenius-i-vanisher.com/ddrsig/2647.png?t=1235917793

Post #19 · Posted at 2012-09-16 09:13:05am 12.8 years ago

Offline Kyzentun
Kyzentun Avatar Member
3,205 帖子
United States
Reg. 2008-02-20

"I'm honestly pissed off."
For a tournament between serious players where accuracy is the most important thing, use Good count. Whover can get the most Goods and still pass the song wins. However much they have of other judgements doesn't matter.

For matches, 2/3 seems best.
silenttype01: Kyzentun is never harsh. He says it how it is.

GENERATION 24: The first time you see this, copy it into your sig on any forum and add 1 to the generation. Social experiment.

Post #20 · Posted at 2012-09-16 12:27:56pm 12.8 years ago

Offline XmatthewX
XmatthewX Avatar Member
612 帖子
United States
Reg. 2011-07-21

How would one go if the people play Standard and Heavy in a Tournament on a DDR Extreme? Points wont work because of extremes system. I was thinking more of a percent like:

Perfect +3
Great +2
Good/O.K +1
Almost +0
Boo/NG -1

Example

Perfect 60 which is 180 pts.
Great 20 which is 40 pts.
Good 10 which is 10 pts.
Almost 5 which is 0 pts.
Boo 5 which is -5 pts.
O.k 10 which is 10 pts.
NG 5 which is -5 pts.

So in total this person got 230 pts.

But my issue is how would I make it fair if someone in Standard can actually beat Heavy and reverse in a fair situation. because some people can play Standard because they so good at the game that they can get all perfects (AAA) and other people are not as good. How can i balance this?
註冊賬號 登錄 Back To Forums

0 User(s) Viewing This Thread (Past 15 Minutes)

©2006-2025 Zenius -I- vanisher.com -5th style- IIPrivacy Policy
Web Server: 4% · Database: 4% · Server Time: 2025-07-08 05:50:58
This page took 0.016 seconds to execute.
Theme: starlight · Language: chinesetraditional
Reset Theme & Language