CPU jumpers

[pepsiman89]

I don´t wnat to registrate me

Ohter options ?

try www.speedy.sh you dont have to register there

[h2min]

I don´t wnat to registrate me

Ohter options ?

http://www.speedyshare.com/

[he10th3r310]

I guess it's not possible to see CPU players jumping at all? Just their results?

[Jussi Koskela]

I guess it's not possible to see CPU players jumping at all? Just their results?

In the current approach is not possible to see the jumps.

[Martin Kafka]

I messed around with the CPU skills a little and I'm not convinced. CPU jumpers with 1600 in Large Hill equal someone who has an online rating of around 1000...

Ouch. If this turns out to be true, it will be very unfortunate. CPU players that are not competitive enough even on the highest level are useless. :/

[Bartek Palewicz]

I guess it's not possible to see CPU players jumping at all? Just their results?

In the current approach is not possible to see the jumps.

LOL, just wondering which leg does CPU jumpers use for telemark.

[Jussi Koskela]

I messed around with the CPU skills a little and I'm not convinced. CPU jumpers with 1600 in Large Hill equal someone who has an online rating of around 1000...

The probability distribution of the jumpers' performance is based on the jumps from the online game, so this shouldn't happen. Even the best jumpers can fail in single competitions but in the long run their skill should match that of the online players.

[Austrian_Skijumpfan_94]

Jeah i´ll see it
One time a player with 600 rating in singleplayer was in the top ten (but with 216 players)

[Risto Salonen]

The probability distribution of the jumpers' performance is based on the jumps from the online game, so this shouldn't happen. Even the best jumpers can fail in single competitions but in the long run their skill should match that of the online players.

In that case there's a bug because in overall competition (10 races) I destroyed the CPU jumper who had 1600 (remember this is large hills only). I also had 15 CPU players with a skill of 1000; the one with 1600 finished second but not by as clear margin as the difference in skill would suggest.

Need to create a more varied testing environment with more jumpers. Could be that it's related to something like the amount of competitors again...

[Jussi Koskela]

If there's any bug I'll be happy to fix it

[Austrian_Skijumpfan_94]

http://speedy.sh/uv9rZ/Ski-Jumpers.rar

PAc0k with with about 220 jumpers

[h2min]

http://speedy.sh/uv9rZ/Ski-Jumpers.rar

PAc0k with with about 220 jumpers

Thank you so much.

[greex1]

all jumpers from the season 11/12:
http://speedy.sh/KgcEw/Ski-Jumpers.rar
lack of real clothes.

[pepsiman89]

it also seems that the CPU jumpers fall way too often

[Jussi Koskela]

it also seems that the CPU jumpers fall way too often

The fall rate is based on the stats from the online game (length, skill and wind are taken into account). I want that the players have equal chances to beat the CPU players even if it resulted unrealistic amount of falls.

One problem could be that the online players were pursuing only the maximum length in the qualification round if they know that they will qualify and this raises the percentage of falls. (I am not differentiating the different rounds any way when I analyze the stats from the online game.)