Scoring System FAQ (cutoff calculation explained!)

**pata70** · Joined: 23 Mar 2007 Posts: 426 Location: Canada

Let's put our minds to it and try to get to the bottom of this!

There are three different scoring methods we have to figure out:

A) The points awarded by the game as shown in your score
B) The base score used by the game to calculate star ratings
C) The points calculated by Ghex (0.4.3) in the note charts

I'm especially motivated to understand the difference between B) and C) so that we can estimate cutoff scores.

From your post you say that points awarded and base score don't match. This is reinforced by the Top Scores last digits.

But as I've pointed out in another thread
http://www.scorehero.com/forum/viewtopic.php?p=179462#179462
it looks like base score and Ghex note charts points also don't match.

For GH2-360 Yes We Can (Expert), the Ghex 0.4.3 note chart points total is 61395 (holds at 149). The proven cutoff is still unknown. The top scores all end in 00.

For GH2-360 Yes We Can (Easy), the Ghex 0.4.3 note chart points total is 24145 (holds at 74). The proven 4-star cutoff is 48390 (base score 24195 -- a difference of 50 points with Ghex). The top scores all end in 50.

I'm not sure how to go forward in this investigation...
_________________

Is life? No. Demorez.

**Barfo** · Joined: 10 Oct 2006 Posts: 2596

well i mean YWC expert, its obvious that the notechart didnt change from PS2 to 360, so im sure that the cutoffs is the same as the GHex one (or at a minimum the same in the holds section which is where the base score difference and actual score difference are), since the GHex agrees with the PS2 cutoff. Also since the top scores for it are still multiples of 50, then its clear that the 360 behaves just like the PS2 and it awards one extra tick of sustain in 5 of the 6 sustains compared to both the GHex calc and the base score. That is basically solved as much as it can be for this song, even though we may not like the implication of the results.

The YWC is a very easy song to proof 4* on anyways, you just have to break like every 4x and then in the 6 sustains hit those at 1x and if you dont get the right mod 50 after that then start over (practice the ending a few times until you are comfortable you can do it every time).

For easy it sounds like the GHex is just missing a note somewhere (thus the 50 off). Thats not even related to this issue, i would guess. And plus im sure there are other cases where GHex doesnt match the rounding of sustains even though im many cases the new one seem to catch some errors. Thats probably the explanation for that song where you had 80 base score difference.
_________________

Watching her, these things she said / "Time," she cried "failed to wait, this time"
***
Hush now / Let it go now / I know it's time to go / Time to let this fall / From my hands

**pata70** · Joined: 23 Mar 2007 Posts: 426 Location: Canada

I'm confused...

Even on Yes We Can (Expert), the GHex total doesn't match the proven PS2 cutoff:

GHex note total for 360 is 61395
GHex note total for PS2 is 61349 +4 hold points and -50 for missing last note? (YB instead of GYB)
PS2 proven base score is 61390

(Are the GHex PS2 note charts linked in the forums made using the same version 0.4.3?)
_________________

Is life? No. Demorez.

**Barfo** · Joined: 10 Oct 2006 Posts: 2596

Yeah i think those are with an older version. That one used to cut off the lowest note of the chords if the song ended on a non-sustain chord, and it had some problems with runs of the same hold notes running into each other, not showing all of them. And i think based on what i saw in the JtF charts direct comparison that i did a while back to chase down that PAL issue that it was handlign rounding differently - compare PS2 to 360 on JtF and some holds are shorter points on PS2 (?). kata would know the details better. If you compare the PS2 and the 360 notecharts you ee that the PS2 one is +1 points for each of the four final sustains. So those two known ghex issues explain (as you point out) the +4 and the -50 difference between those scoresof the two charts. The ghex estimate for 360 is more reliable than the older PS2 notechart ones which is why i always quote those (for songs that i dont think the notecharts).

So thinkign only in terms of the 360 chart, the ghex disagrees with the actual gameplay in 5 of those sustains being +1 in actual gameplay (ie they are worth 150 and not 149). Whereas in the base score calculation from the game, the proven base score shows that it is counting them as 148. Funky business. I didnt notice it on expert that the ghex was not doign -2 on the chords like the base score obviously does. on the other diffs where its single note sustains the ghex and the game base score agree and are counting those at -1 compared to the gameplay.

What this means is that JCirri was right in requiring that cutoffs only get entered based on proven pics, as there will always be some doubt between whether we (ghex) are calculating the base score right in terms of the gameplay, and even if we get it right in terms of gameplay (or even if on easy somebody was able to do a 100% at 1x with all hold points run and verify the actual base gameplay value) we now know that at least in some cases the game's own chart calculation of base score can disagree (slightly) with the actual gameplay base score.
_________________

Watching her, these things she said / "Time," she cried "failed to wait, this time"
***
Hush now / Let it go now / I know it's time to go / Time to let this fall / From my hands

**pata70** · Joined: 23 Mar 2007 Posts: 426 Location: Canada

Hmm. Could it be as simple as:

Both gameplay scoring and GHex round the sustain points up or down, maybe differently, and internal base score calculation doesn't round at all (for each individual sustain), but only rounds the total base score?

I wonder if the makers of GHex could try to accumulate the base score including any fractions of sustain notes and show the total? If it hasn't been done already...

Also, if GHex shows note scores for each measure that don't match the gameplay scoring, and don't match the internal base score either, then should it be considered incorrect? Maybe we could get rid of this "third" scoring method if we have it match exactly with one of the other two... and in this case, which one would we want to see in the note charts, the internal base score or the gameplay scoring?

Sorry about the rambling, if this doesn't make sense for GHex or has already been discussed before, I really haven't been following GHex development and can't try it because I only have the 360 version without the secret recipe

_________________

Is life? No. Demorez.

**debr5836** · Joined: 11 Apr 2007 Posts: 1159 Location: Fort Collins, CO

I'm working on a GH midi-parsing project of my own.

I think the sustain tick errors lay in where you put the rounding. In geek speek, its the difference between

sustain_score = multiplier * chord_size * int ( 25.0 * note_beats + 0.5 + 1e-7)

vs.

sustain_score = multiplier * int ( 25.0 * chord_size * note_beats + 0.5 + 1e-7)

vs.

sustain_score = int ( 25.0 * multiplier * chord_size * note_beats + 0.5 + 1e-7)

I'm guessing that GHEx is using the 2nd choice (thus getting odd number results from two note chords), while GH2 is using the top choice. I've had much better luck in aligning to reality with some variant of the top equation.

Perhaps we can bribe kata into moving the rounding around in the next GHEx to see if he can line up the calculated base scores a little more tightly.

**ChainsawSolo** · Joined: 17 Mar 2007 Posts: 78 Location: Madison, WI

**Barfo** · Joined: 10 Oct 2006 Posts: 2596

**PenguinMan** · Joined: 04 Mar 2007 Posts: 324 Location: Galesville, WI

wow, how long did this take?

_________________

Goals:
Get the Log GH2: [6 songs to go]
GH 80's [2 songs left]
Beat Jordan (expert) [59%]
Electric Eye: Guitar Solo E

**debr5836** · Joined: 11 Apr 2007 Posts: 1159 Location: Fort Collins, CO

Alright, I think I've had a bit of a breakthrough on calculating cutoffs.

Let's considere pata's breakdown (or at least a similar one)

(a) GH2 Gameplay score calculation
(b) GHEx score calculation
(c) Base

I'm hypothesizing that the sustain formulae for each of these cases is as follows:

(a) int ( 25 * chordsize * <beats> + 0.5) * multiplier
(b) int ( 25 * chordsize * <beats> ) * multiplier
(c) int ( 25 * <beats> ) * chordsize * multiplier

In running this through Surrender/X360/Expert, I get the following three "base scores"

(a) 66226
(b) 66157 (same as GHex)
(c) 66051

Note that 2 x (c) corresponds exactly to the 4* cutoff. This could be coincidence, but I'm hoping now.

I've got some other bugs in here, but I thought I would share with you the glimmer of encouraging news here.

I'm hoping that if we can demonstrate an algorithm that aligns with all current known proofs, we can convince JCirri to populate all of the holes with at least "red" numbers.
_________________

Check out my SP Galleries for GH1-PS2, GH2-PS2, GH2-X360, GHRT80S-PS2, and GH3
GH2 Log: 9/2/2007

**Slessie** · Joined: 08 Jul 2007 Posts: 6

Sweet guide, thanks.

**Gibertram** · Joined: 16 Jun 2007 Posts: 180 Location: Derry, NH

interesting... Has any song ever been 9*ed?

**LocalH** · Joined: 30 Oct 2006 Posts: 1400 Location: MiloHax

Ok, I'm thinking after doing some talking with a few people that the values given here are slightly off in terms of whammying additional SP. JC says:

**MuDvAyNe** · Joined: 08 Jan 2007 Posts: 1577 Location: St. Peters, MO

I will second that you can get an activation out of the third phrase in The Warrior...
_________________

Just click, you won't regret it, I promise.

someone who clicked the button wrote:

Wow, I do not regret clicking that button. Awesome Videos!

<--Click the banner for the video

**Apoc112** · Joined: 05 Aug 2007 Posts: 10 Location: Erie, PA, USA

thank you for flooding my mind with GHII knowledge while i'm at work... it really helped distract me from my (potential) customers.