Cat Physics [Donut Games] (v1.11 - 10 new challenges)

[Pfadfinder]

Does anyone have an idea how the counting system works?

I only figured out/assume:

- it counts back
- it starts with 1000
- it stops when the tail of target is reached
- not time is counted but length of path
- dotted line represents path travelled
- line does not show any tiny move but smoothens the graphical representation of path
- decisive is actual path not line representing path
- teleported way does not count
- center of ball is decisive
- rotation of ball does not matter
- end position in closed tail is decisive not the point where ball touches tail first


Wiz mentioned lately that sometimes a longer path (line???) brings more points. I assume that this an illusion because of the smoothing effect of the graphical representation of the actual path. however that assumption needs further observation.

I wonder how the length of the path is
measured. I doubt that the physic engine includes differential calculus methods. Could it be that every Pixel crossed by the center of the ball deducts 1 point from 1000? If this is true than 248 should be between the starting and end position of the center of the ball in level 1. Has anyone counted already?

[Cowboy.Henk]

Originally Posted by Pfadfinder

Does anyone have an idea how the counting system works?

I only figured out/assume:

- it counts back
- it starts with 1000
- it stops when the tail of target is reached
- not time is counted but length of path
- dotted line represents path travelled
- line does not show any tiny move but smoothens the representation of path
- decisive is actual path not line representing path
- teleported way does not count
- center of ball is decisive
- rotation of ball does not matter
- end position in closed tail is decisive not the point where ball touches tail first


Wiz mentioned lately that sometimes a longer path (line???) brings more points. I assume that this an illusion because of the smoothing effect of the graphical representation of the actual path. however that assumption needs further observation.

I wonder how the length of the path is
messuared. I doubt that the physic engine includes differential calculus methods. Could it be that every Pixel crossed by the center of the ball deducts 1 point from 1000? If this is true than 248 should be between the starting and end position of the center of the ball in level 1. Has anyone counted already?

Hi pfad, in lvl 10 I have had paths that are clearly shorter than the max with a lower score. I agree that distance is likely not computed according to Euclidean metrics.

[Pfadfinder]

Originally Posted by Cowboy.Henk

Hi pfad, in lvl 10 I have had paths that are clearly shorter than the max with a lower score. I agree that distance is likely not computed according to Euclidean metrics.

Do you have screenshots for further analysis?

[Pfadfinder]

Some findings and further assumptions

[Pfadfinder]

This is an overlay of level 1 start and target reached.

You will notice:

- that ball in tails is higher than on ground level
- path is slightly curved at beginning and end
- perimeter of ball is one pixel lower than ground level (well, that is noticable in picture in part 2)
- the target cat is one pixel lower than the start cat (I do not assume that this is intention.)

I measured that distance between centres of ball is approx. 325 pixel.

The counting stops at 752. Therefore 248 is deducted from 1.000. Therefore there is a contradiction to my assumption in part 1 that every crossed pixel deducts 1 point.

I therefore adjust my theory as follows:

The counting system counts every pixel crossed. The sum is multiplied by a constant. The constant is below 1.

The constant may be dereived as follows:
325 direct line from centre to centre of balls
248 = 1000-752

325:248=1:x
x=248/325= 0,7630769230769231

0,76 and so on is to odd to be choosen by Ola. I assume therefore that Ola chose a nice constant as 0,75 = 3/4.

248*4/3= 331,66 ~ 332

I assume that the slight curve makes the path longer than the direct connection.

Taking the second chart in part 2 into account you may add easily 6-7 pixel to the measured direct line of 325 and come to 332. And with an adjustment constant of 3/4 you come easily to 248 and finally 752.

So my thesis is that the counting system uses the following formula:

score = 1000 - (number of crossed pixel * 0,75)

It will not be easy to proof the formula, though.

I did not believe that an analysis of level 1 provides so much remarkable observations.

You may ask what is the purpose of this analysis. Well, for the time beeing it is only basic research. Maybe the findings/assumptions lead to something later.

You may ask why should Ola chose a constant of 0,75. Well, I have no clue.

Maybe the whole idea is bullshit after all. But anyhow, does anyone have a better explanation of the counting system?

[Cowboy.Henk]

The distance computation is based on the original iPhone 3 screens, not on the retina display. This could cause the factor of 0.76...

[Cowboy.Henk]

Originally Posted by Pfadfinder

Do you have screenshots for further analysis?

Nope, I did not hold on to these ss...

[cy@n]

Originally Posted by Pfadfinder

[IMG]http://forums.toucharcad[...]But anyhow, does anyone have a better explanation of the counting system?

There was an information in this huuuuuuge forum thread, you might be interested in:
http://forums.toucharcade.com/showpo...postcount=9656

[Cowboy.Henk]

Originally Posted by cy@n

There was an information in this huuuuuuge forum thread, you might be interested in:
http://forums.toucharcade.com/showpo...postcount=9656

Thx cy@n, that was the post I was thinking of.

[Pfadfinder]

Originally Posted by cy@n

There was an information in this huuuuuuge forum thread, you might be interested in:
http://forums.toucharcade.com/showpo...postcount=9656

yes this very interesting. to my mind it suggests that the counting systems of cp on iphone 3 and 4 are identical

thank u both for feedback discussion and contribution