View previous topic :: View next topic |
Author |
Message |
Kinitawowi
Joined: 18 Sep 2007 Posts: 1074 Location: Newham, London, UK
|
Posted: Thu Jan 10, 2008 3:39 pm Post subject: |
|
|
Venge wrote: | If the 5* multiplier is 2.8x (40% more than 4*) instead of 3x (50% more) then shouldn't 6* be 3.92x, 7* = 5.488x, 8* = 7.6832x, 9* = 10.75648x? Or working off the GH1:GH2 ratio: 5* = 3.0:2.8 = 15:14 = 93 1/3%. Which would make 6* = 3.72x, 7* = 4.65x, 8* = 5.58x, 9* = 6.51x. |
Well, the ratio system is right out, because it gets the 4* wrong - that's fixed at 2x, and your calculation (GH1:GH2 = 15:14 -> 2:x) results in x = 1.86666x - and the game knows that that should be 2x as well. As for the exponential approach...
Just no. _________________
My Crapchievements Thread GOAL: The Log [Beast And The Harlot 4.7*]
"Our Father, who art in Heaven... please, stay there." - Saint Etienne, New Thing |
|
Back to top |
|
|
pk217doc
Joined: 08 Nov 2006 Posts: 145
|
Posted: Thu Jan 10, 2008 3:43 pm Post subject: |
|
|
Venge wrote: | JCirri wrote: | Code: |
| | Avg. Mult. | SH Decimal Rating |
| In-game Rating | GH1 | GH2 | GH1 | GH2 |
------------------------------------------------------------
| 4-stars | 2.0x | 2.0x | 4.0 | 4.0 |
| 5-stars | 3.0x | 2.8x | 5.0 | 5.0 |
| 6-stars | 4.0x | 3.6x | 6.0 | 6.0 |
| 7-stars | 5.0x | 4.4x | 7.0 | 7.0 |
| 8-stars | 6.0x | 5.2x | 8.0 | 8.0 |
| 9-stars | 7.0x | 6.0x | 9.0 | 9.0 |
|
|
After reading the 1/8/08 updates in tma's Song Scores, Extended Star Ratings and Cut-Offs FAQ 20080108 it struck me that the math on the GH2/GH3 extended ratings might be faulty. I searched that thread and this one (the only logic places) to see if this has been brought up, so if there's an errant post out there, sorry, but it really should be addressed, either way, in this post.
If the 5* multiplier is 2.8x (40% more than 4*) instead of 3x (50% more) then shouldn't 6* be 3.92x, 7* = 5.488x, 8* = 7.6832x, 9* = 10.75648x? Or working off the GH1:GH2 ratio: 5* = 3.0:2.8 = 15:14 = 93 1/3%. Which would make 6* = 3.72x, 7* = 4.65x, 8* = 5.58x, 9* = 6.51x.
I understand the extended scoring is an arbitrary system we made up ("we" = the "royal" we = JCirri) and that logic doesn't necessarily apply. I understand this is the way it's been done for quite a while now and changing it would upset those that are opposed to change, not matter the reason. I understand this makes the extended scores more difficult and changing it would upset those that are above the old cutoffs, but below the "logical" ones; especially the 8*'s, since the first set of calculations put the 8* higher than the old 9*, which has yet to be achieved (?). I'm not even proposing that we actually change it, I was just curious if this was considered and if there was a reason it was rejected. |
First off, I think you've put too much thought into all of this, since the extended star ratings is very simple. For GH1, they figured that since the difference between the multipliers of 4-star and 5-star was simply 1x, that a 6-star rating would simply be 1x higher that 5-star, 7-star would be 1x higher than 6-star and so on.
Same goes for GH2 and beyond, since the difference this time between 4 and 5-star was 0.8x, they added 0.8x to each extended rating to get the next (i.e. 2.0x, 2.8x, 3.6x, 4.4x, 5.2x, 6.0x). Or basically the formula could be 2+(a-4)*0.8, where "a" is the star rating you're trying to get.
Now, if you applied the logic that you did on GH2 to get the 3.92x, 5,488x, 7.6832x, and 10.75648x multipliers (adding 40% in other words) for extended ratings, and did the same thing for GH1, you would have to add 50% to each higher rating, giving a series of multipliers to be 2.0x, 3.0x, 4.5x, 6.75x, 10.125x, and 15.1875x. This would obviously make extended ratings for GH1 suck, since the highest rating anyone would probably ever get using these multipliers would be 6-star.
As far as trying to create a GH1 to GH2 ratio (the 15/14 that you mentioned), you really can't do that since the scoring mechanisms are different for those games. And you wouldn't end up getting the 2x multiplier for a 4-star on both games that way. _________________
Alakaiser wrote: | They should be around some time between now and when they're up. |
|
|
Back to top |
|
|
pk217doc
Joined: 08 Nov 2006 Posts: 145
|
Posted: Thu Jan 10, 2008 3:49 pm Post subject: |
|
|
Kinitawowi wrote: | Venge wrote: | If the 5* multiplier is 2.8x (40% more than 4*) instead of 3x (50% more) then shouldn't 6* be 3.92x, 7* = 5.488x, 8* = 7.6832x, 9* = 10.75648x? Or working off the GH1:GH2 ratio: 5* = 3.0:2.8 = 15:14 = 93 1/3%. Which would make 6* = 3.72x, 7* = 4.65x, 8* = 5.58x, 9* = 6.51x. |
Well, the ratio system is right out, because it gets the 4* wrong - that's fixed at 2x, and your calculation (GH1:GH2 = 15:14 -> 2:x) results in x = 1.86666x - and the game knows that that should be 2x as well. As for the exponential approach...
Just no. |
Ninja'd on the 2x multiplier not being the same, but Venge never mentioned anything about an exponential increase in the multiplier. The 40% increase is strictly multiplication and/or addition (depending on how you do it). So I'm not entirely sure how you were able to get an exponential graph, unless I have totally forgotten some of my math. _________________
Alakaiser wrote: | They should be around some time between now and when they're up. |
|
|
Back to top |
|
|
Kinitawowi
Joined: 18 Sep 2007 Posts: 1074 Location: Newham, London, UK
|
|
Back to top |
|
|
pk217doc
Joined: 08 Nov 2006 Posts: 145
|
Posted: Thu Jan 10, 2008 4:17 pm Post subject: |
|
|
Kinitawowi wrote: | pk217doc wrote: | Ninja'd on the 2x multiplier not being the same, but Venge never mentioned anything about an exponential increase in the multiplier. The 40% increase is strictly multiplication and/or addition (depending on how you do it). So I'm not entirely sure how you were able to get an exponential graph, unless I have totally forgotten some of my math. |
multiplier = 2 * [ 1.4 ^ (rating - 4) ]. Go back and re-read the bit in your maths textbooks about compound interest.
EDIT: relabelled equation |
Oh, I see where I went wrong on that thought. My bad on that part of it. I was just thinking of it the wrong way when I looked at your graph I guess. It's been a while since I've even had to think about any kind of math problems like this, but that's not much of an excuse to forget something as simple as compound interest.
But yeah, adding the 40% to each successive multiplier would give you the formula Kinitawowi mentioned, and in turn that would kill the usefulness of the extended star ratings. _________________
Alakaiser wrote: | They should be around some time between now and when they're up. |
|
|
Back to top |
|
|
Venge
Joined: 09 Aug 2007 Posts: 254 Location: Yuma, Arizona
|
Posted: Thu Jan 10, 2008 4:34 pm Post subject: |
|
|
pk217doc wrote: | First off, I think you've put too much thought into all of this, since the extended star ratings is very simple. |
Problem is, I didn't put hardly any thought into it. As I said, I was reading tma's faq and it occured to me that just adding .8 to each sequential star ignores the fact that if the rating were reversed 3* would be 1.2x base score, which implies just adding .8 isn't 100% accurate. The two examples I could think of off the top of my head suffer the same problem (3- and 4- stars don't match up), but I didn't think about it that far -- which now makes me feel really stupid because they're flawed in exactly the same way I was pointing out .
The extended star ratings is very simple because we've made it that way; we just made some shit up and said that's the way it is. The math behind it is not very simple. I'm am not a math professor, but I know there are some very learned "mathies" here and wanted to make sure it was considered. JCirri's graph pretty much put it in perspective though: it might not be 100% accurate, but it's close enough.
EDIT1: Why the Hell was there a "[/i]" at the end of my post?!?
EDIT2: ^^ Rhetorical question ^^ |
|
Back to top |
|
|
Kinitawowi
Joined: 18 Sep 2007 Posts: 1074 Location: Newham, London, UK
|
Posted: Thu Jan 10, 2008 4:45 pm Post subject: |
|
|
Venge wrote: | Problem is, I didn't put hardly any thought into it. As I said, I was reading tma's faq and it occured to me that just adding .8 to each sequential star ignores the fact that if the rating were reversed 3* would be 1.2x base score, which implies just adding .8 isn't 100% accurate. The two examples I could think of off the top of my head suffer the same problem (3- and 4- stars don't match up), but I didn't think about it that far -- which now makes me feel really stupid because they're flawed in exactly the same way I was pointing out . |
Oh, I've realised that myself a long time ago - the spreadsheet I use on my own computer has some special case fudging to fix 3-stars at 1.0x and then automatically works out 4-stars at 2.0x, 5-stars at 2.8x, 6-stars at 3.6x etc. Hypothetically, as far as the game's concerned it's irrelevant. The game says "if you get 2.8x you get 5-stars, otherwise if you get 2.0x you get 4-stars, otherwise you get 3-stars". That's why the 3-star calculations can be "wrong" under any model (because the game doesn't hold a value for them), but the 4-star and 5-star ones have to be right.
Quote: | The extended star ratings is very simple because we've made it that way; we just made some shit up and said that's the way it is. |
Probably the most accurate thing said in this thread. _________________
My Crapchievements Thread GOAL: The Log [Beast And The Harlot 4.7*]
"Our Father, who art in Heaven... please, stay there." - Saint Etienne, New Thing |
|
Back to top |
|
|
bulletformyjew
Joined: 13 Feb 2008 Posts: 223 Location: Manchester UK
|
Posted: Mon Mar 10, 2008 2:29 pm Post subject: |
|
|
on co-op Gh3 we got a times multiplyer on 2.710 but it was 5*????
Why? |
|
Back to top |
|
|
tma
Joined: 03 May 2007 Posts: 1414 Location: Australia
|
Posted: Mon Mar 10, 2008 8:41 pm Post subject: |
|
|
bulletformyjew wrote: | on co-op Gh3 we got a times multiplyer on 2.710 but it was 5*????
Why? |
We'd all like to know that. |
|
Back to top |
|
|
CovyKiller
Joined: 10 May 2008 Posts: 127
|
Posted: Thu May 29, 2008 7:32 pm Post subject: |
|
|
Let me just understand one thing, it is completely impossible to achieve a nine-star rating on any song, correct?
PS: didn't really sift through all the pages, should've. |
|
Back to top |
|
|
blob
Joined: 11 May 2007 Posts: 2177 Location: Dallas, TX
|
Posted: Thu May 29, 2008 7:35 pm Post subject: |
|
|
Its not impossible in the world of GH. But no song at this moment in any game can get 9*. _________________
|
|
Back to top |
|
|
thecaptainof
Joined: 04 May 2007 Posts: 7571 Location: ¯\(°_o)/¯
|
Posted: Thu May 29, 2008 7:43 pm Post subject: |
|
|
blob wrote: | Its not impossible in the world of GH. But no song at this moment in any game can get 9*. |
...except The Seeker X/X co-op. _________________
yksi-kaksi-kolme wrote: | Wow Mr. Mad, who fucked your buffalo? |
|
|
Back to top |
|
|
zsjostrom35
Joined: 17 Feb 2008 Posts: 2130 Location: Columbus, Ohio
|
Posted: Sat May 31, 2008 4:03 am Post subject: |
|
|
thecaptainof wrote: | blob wrote: | Its not impossible in the world of GH. But no song at this moment in any game can get 9*. |
...except The Seeker X/X co-op. |
Well, nobody knows exactly how the co-op star ratings work, so you can't really say that they got nine stars. Their average multiplier was indeed over 6.0, however, which is why most people refer to it as a nine-star run. _________________
|
|
Back to top |
|
|
galvao1090
Joined: 09 Apr 2007 Posts: 26 Location: Brazil
|
Posted: Sun Sep 07, 2008 1:37 pm Post subject: |
|
|
Just a suggestion to the site:
Why don't you extend the concept of number of stars to the total score in each game? Then the users can be compared by their "Star Rating" too... (For example, a gamer that achieved an average multiplier of 3,6 across all songs a 6-star player) _________________
League 10 AA/Hard 1st Place! Woo Hoo!
Vocals career stats
[58/58] Rock Band 1
[84/84] Rock Band 2
[83/83] Rock Band 3
[1/44] Green Day: Rock Band (I only played Boulevard of Broken Dreams and don't have this game)
[30/44] The Beatles: Rock Band
[45/45] Lego: Rock Band
[8/18-1] Rock Band: AC/DC Live (18-1 because Jailbreak seems to be broken on PS2) |
|
Back to top |
|
|
liljj93
Joined: 12 Sep 2008 Posts: 77
|
Posted: Thu Sep 18, 2008 1:11 am Post subject: |
|
|
i didnt know a 6.0 avg. multiplier was possibly |
|
Back to top |
|
|
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
Copyright © 2006-2024 ScoreHero, LLC
|
Powered by phpBB
|