Scoring of building

[fredjust]

hello

I always think that building wasn\'t enough reward in block Out

So I compare Tetris and BlockOut
and there is the result :



But in Block Out you need some luck to make a 5 lines so it\'s no comparable


The biggest difference is in drop point

They take a huge place in BlockOut
and nothing in tetris<br><br>Post edited by: fredjust, at: 2007/07/24 21:31

[Herc]

i agree: building higher stacks and removing more layers at once should be worth more.

here my suggestion:

my most simple and intuitive score formula:

1 cube = 10 points

--> one layer in 5x5 = 250 points

score = deleted_cubes*10*deleted_layers

i.e. only cubes that are removed from the pit are worth something. and score is multiplied by number of simultaneously deleted layers:
1 layer = 250 points
2 layers = 1000 points
3 layers = 2250 points
4 layers = 4000 points
5 layers = 6250 points.

so - this is exponential and logic: the only score is awarded to removing cubes from the pit. no score for dropping a block. is it worth giving points to something you just dropped? i think no. also, i think it is not necessary to give higher scores for early drops, if the gameplay is changed into the following:

1. method: each game lasts exactly 10 minutes. (or rather 5 minutes, because for most people 10minutes is just too long) to score high, you have to play as fast as possible.

2. method: you can play at your own speed, but score is afterwards scaled depending on you overall game time. if you took 20 minutes, your score will be divided by 2.
if you needed only 5 minutes till you drop out in skull-level, your score will be multiplied by 2. (if you die in an earlier level, the multiplier must be adapted somehow)

i like method one more. its more like the short and fun 180seconds mode in normal tetris.

but the problem is: JeanLuc wants to be compatible to original blockout. so i think he will not change scoring.

or maybe an optional, \"modern\" scoring scheme? the same as the options with look-and-feel (sound and marbe stuff) ? the problem with this is: this will make highscore competition less interesting, because some will play classic blockout scoring, some will play modern scoring... i think we currently do not have enough players to have 2 simultaneous highscore lists...

[fredjust]

hello

your scoring is the same than actual Block Out
because I calculate the points/ligne score

so

>1 layer = 250 points
>2 layers = 1000 points
>3 layers = 2250 points
>4 layers = 4000 points
>5 layers = 6250 points.

=>

250/1=250
1000/2=500
2250/3=750
4000/4=1000
6250/5=1250

it\'s linear

in Tetris \"Points/ligne\" is an exponetial function
and not \"points for x line\"

>no score for dropping

remove only this and gaming will completly change

a lot of test is necesary to change scoring in the good direction

>1 method: each game lasts exactly 10 minutes

this is an another game mode like in tetris DX on GBC

- marathon
- 40 lines
- ultra (3 minutes)

>but the problem is: JeanLuc wants to be
>compatible to original blockout. so i think he
>will not change scoring.

this change is for BlockOut 3 ;)

[fredjust]

in fact my first graph is :

\"point per line for x line\"

100 for one line if you do one line
190 for one line if you do two line (so 190*2 points)
...
450 for one line if you do five line

I\'m interesting to see the hight score table without drop point ;)

is the order will change ?

[fredjust]

> I\'m interesting to see the hight score table without drop point

the 15 first score in a new order without drop point :



...<br><br>Post edited by: fredjust, at: 2007/07/26 05:37

[Herc]

very interesting! fine table. i will think about it.

i still want to have a game mode, that is a cross-over between marathon and 3minutes and 40 lines. that has all the good things of all these modes.

40lines: try to play as fast as possible, but only time counts.
marathon: speed is not important, only good stacking, tetrises and later, survival.
ultra: is already a nice mix between pure speed and pure score attack. in ultra, you must play fast AND good to get high scores.

the stupid thing is, that in tetris dx you can select a level (drop speed) even in ultra mode. so if you select level 9, you get much higher scores, as if you select level 1. if you play really fast, starting in level 1, you will get much less score as if you start in level 9 and play equally fast and good.

so i would
1. remove different scores with different level (in ultra mode)
2. remove drop point scores (because they are completely useless in ultra mode)

and i have another idea: instead of linear / exponential reward for making lines /layers, why not just punish the case making single lines /layers? and punishing by blocking user input during the time the line blinks and gets deleted. (i.e. line clear delay)

1 layers at once: 1.0 seconds line clear delay
2 layers at once: 0.5 sec line clear delay
3 layers at once: 0.25
4 layers at once: 0.0

so the gamer can play fastest if removing always 4layers. if only removing single layers, he will loose lot of important time in ultra mode.

on the other hand, rewarding is always more positive than punishing ...

[jlp_38]

Hi,

Very good ideas B). I think that adding new modes can greatly improve the game. But I won\'t work on this these next days. I\'m currently porting Blockout II to Linux platform. I think this can brings new gamers.

@Fred:

Good idea for the score page. I will add this. I think during the week-end as it is not a big job.

[Herc]

great! a linux port will definitely bring many new gamers! will you use OpenGL? OpenAL? SDL ?

[jlp_38]

I chose the SDL/OpenGL solution. I already made few test and it looks promising.
I\'ll keep you informed.

[Herc]

thats cool, because using open standards is always better...
about 10 years ago i first tried directx (was it 6.0 - i dont remember), and was frustrated about the strange coding style and hard to read and understand code. (may have changed with modern directx versions).

then i tried OpenGL together with libfltk and had immideate success in doing some basic 3d stuff. it was a real enlightenment and fun to code with OpenGL.

i only lateron discovered SDL - maybe around 2001. it made keyboard input and joystick stuff so much easier. in fact, i never tried to understand the microsoft api - it always looked cryptic to me. again, with SDL, everything seemed so much more easy.

and a nice side effect was that i automatically gained cross platform compatibility.

so - i really appreciate your decision to use SDL and OpenGL - definitely i right decision in my eyes. and - who knows how long microsoft will be on market - after its flawed vista.... who knows - maybe in 10 years linux or MacOS (sort of linux too) dominate microsoft.

the only good thing i can say about microsoft is the visual studio ide. thats a great and productive thing. love it. much faster than gcc in my eyes. and easier. especially together with www.visualassist.com. jean luc, if you do not know this fantastic tool, you must try it. its the best code completion tool ever. much better than intellisense. i cannot live without visualassist any more. and its cheap too, only 50$ if i remember right.

[jlp_38]

Hi,

That\'s true that OpenGL is more intuitive that Direct3D even if the D3D API has been simplified in the last releases. However, I think that Direct3D is still better. It has a more powerfull mathematical API (D3DX) and from my point of view, the global API is more usefull. For instance, in D3D you have 3 matrix (1 for the projection,1 for the camera and 1 for the scene) while you have only 2 in opengl (projection and camera-scene). OK you can always do want you want even with 2 matrix, but playing every time with glPopMAtrix and glPushMatrix is not really useful . There is much small things like this that make D3D better. And last but not least, the Direct3D SDK (including complete documentation and samples) is really great comparing to the OpenGL one.

Anyway, Blockout II is now running with SDL/OpenGL. Graphics and controls works at 100%. It remains the sound and the HTTP stuff. I will do that tomorrow.

[Herc]

hell! how can you learn that stuff so quickly ? or did you know openGL and SDL beforehand ? ok, SDL is quite easy, but a kickstart in openGL in less than a week is nothing else but amazing!

will you stay with openGL and SDL for win32, too? or will you make two versions, linux=sdl+opengl, win32=native win32api, direct3d?

[jlp_38]

I learned OpenGL and SDL during this week. I made my first rotating textured cube last Monday or Tuesday. But when you already know well a 3D API, It is generally simple to learn another one.

I think I will keep 2 versions, one D3D for Windows and one SDL/OpneGL for other platforms. The rendering with Direct3D is slightly better especially with filtered textures.

I started with the HTTP stuff, i looked at libCUrl and wget but unfortunately, any of these APIs support the HTTP PUT method. I think i will modify and use wget.<br><br>Post edited by: jlp_38, at: 2007/07/29 09:29

[fredjust]

I try a lot of thing for scoring :



ORIGINAL : 100 - 370 - 810 - 1420 - 2215
it\'s a proportion to easily compare with other


SQUARE : 100 - 400 - 900 - 1600 - 2500
(100*NbLines^2)
points per ligne is : 100, 200, 300, 400, 500
HERC, it\'s exactly your system (in proportion)

EXPO : 100 - 400 - 1200 - 3200 - 8000
100*2^(NbLines-1)*NbLines
with this system points per lines double for each floor
100 Pts/lines if you do one line
200 Pts/lines if you do two lines (200*2=400)
400 Pts/lines if you do tree lines (400*3=1200)
800 Pts/lines if you do foor lines (800*4=3200)
1600 Pts/lines if you do five lines (1600*5=8000)

I try a MIDDLE solution in the test version

100 - 400 - 1000 - 2000 - 4000

Points per line is 100, 200, 333, 500, 800<br><br>Post edited by: fredjust, at: 2007/07/31 05:48