Deluxe Ski Jump Forum

Hi! This is greex's idea. He wrote this in polish topic and asked someone to translate to english, so i did . About CPU, like topic says.

Instead of setting level of skills and stability of shape we would write rating (like online) of player. We could also pick skills (good, normal, weak) of player on each type of hill (you know, flying, large, normal).
Rating would imitate real players skills with rating we wrote, for example:
Simon Ammann - 1700. So Jussi look at players with rating 1700-1800 (or something else) and check how they jump, how many WOTs they won and other things (he could use dsjtournaments.com). With this he could set how does our Simon Ammann with 1700 rating would jump.

Example of all:
-Marco Grigoli - 900 - Flying hills - very bad, Large Hills - very bad, Normal hills - very bad
-Kamil Stoch - 1600 - Flying hills - good, Large hills - very good, Normal hills - normal
-Robert Kranjec - 1600 - Flying hills - very good, Large hills - good, Normal hills - normal
-Anders Bardal - 1750 - Flying hills - good, Large hills - very good, Normal hills - very good

...and:
-Kazuyoshi Funaki from 1998 - 1900 - Flying hills - very good, Large hills - very good, Normal hills - very good

yeah i also thought about that few weeks ago
would be nice especially because there are different jumper with different skills on each hill

and maybe also kinda one tab for ordinary jumper (for record hunting)
and one for all cpu jumper because it would annoying to change that all the time youre jumping

1750-1900 ? WTF? 1770 Janne Ahonen , 1750 Adam Malysz , Sven Hannawald,Schlieri 1700+ , noone more than that , Robert Kranjec 1600 , Anders Bardal 1600 (he isnt so better than Kranjec,maybe even worser) ... and like that

SuperAdo111 wrote:1750-1900 ? WTF? 1770 Janne Ahonen , 1750 Adam Malysz , Sven Hannawald,Schlieri 1700+ , noone more than that , Robert Kranjec 1600 , Anders Bardal 1600 (he isnt so better than Kranjec,maybe even worser) ... and like that

Whatever, just an example...

love the idea

jop great idea

No in DSJ4. This is very hard to do this. I like system from DSJ3

whatever what rating system cpu players gonna have
we need to set the skill on the different hill types

I don't think putting one parameter is a good idea. In DSJ4 there are 4 things that decide on good jump:

1. Takeoff Timing (means finding best spot at the end of the inrun + pressing both mouse buttons equally)
2. In-Takeoff Positioning (means moving your mouse to find best position at the begining/during takeoff)
3. Flying Technique (means keeping your mouse movement smooth and steady during flight)
4. Landing Technique (means finding the right time to perform telemark/cheaty telemark/both legs landing)

And I think these should be the parameters to set on the CPU plus - fifth parameter, Form.

nexen84 wrote:I don't think putting one parameter is a good idea. In DSJ4 there are 4 things that decide on good jump:

1. Takeoff Timing (means finding best spot at the end of the inrun + pressing both mouse buttons equally)
2. In-Takeoff Positioning (means moving your mouse to find best position at the begining/during takeoff)
3. Flying Technique (means keeping your mouse movement smooth and steady during flight)
4. Landing Technique (means finding the right time to perform telemark/cheaty telemark/both legs landing)

And I think these should be the parameters to set on the CPU plus - fifth parameter, Form.

That's very nice! Maybe just like in DSJ 3 (with %) but not 2, but 5 options.

The skill level adjustment will probably be quite similar to that of DSJ3, but I'll be using ratings like in DSJ4 Online. It's not possible to differentiate the skills in different parts of the jump with the statistical approach. Differentiation based on the hill size can be implemented, though.

nice

Rating = 600-1700 or more/less?

Wow, fantastic news Jussi Thanks

Jussi Koskela wrote:The skill level adjustment will probably be quite similar to that of DSJ3, but I'll be using ratings like in DSJ4 Online. It's not possible to differentiate the skills in different parts of the jump with the statistical approach. Differentiation based on the hill size can be implemented, though.

One question- will we be able to watch CPU jumps?