Nous sommes actuellement le 26 Juin 2022, 07:12


[AI] [0047505] Bug Hillstat

Modérateur: Cyanide Team

  • Auteur
  • Message
Hors-ligne

Organsaft

  • Messages: 2
  • Inscrit le: 21 Juin 2017, 11:34

Re: [AI] Bug Hillstat

Message21 Juin 2017, 11:40

I think everyone would be happy, if you adjust the value 85 to 80 or even 75 or 70 to balance the relation between mo und hill stat. I think this would be the easiest solution. Please think about it!
Hors-ligne

suigfdhuibsyret

  • Messages: 11
  • Inscrit le: 05 Août 2013, 23:25

Re: [AI] Bug Hillstat

Message21 Juin 2017, 11:58

3emeType a écrit:
suigfdhuibsyret a écrit:Oh sorry, i put the videos on private. They should work now. I dont have replays though. I'll think about them next time. Thx for replying


Thx.

What I see here is more an exploit with trains than a problem with hill stat.
It looks like you've found a way of escaping with your team thanks to the train. Once you do that, you use your team to take time, and not your hill stat.
AI should better follow your train.

Another AI problem: on Plouay, Gilbert should have attacked you on the last climb. By doing that he would have won.

And with a correct sprint stat for modified Quintana D. (70 is impossible for a climber that has no punch), you would have lose also. So, in both races, modified Quintana D. would have lose.

So we have work to do to improve the AI :)
Thanks. Very useful videos.


wait, really? ok, I'll try to do the same without making a train, but its pretty hard if the ai doesnt even follow amador of all people on 85 effort on extreme. I'm also pretty sure gilbert couldn't attack on the final hill, u see him standing up on his bike, meaning he's putting an immense effort. Also explains why he cant beat dayer in the sprint with way way better secondary stats.
Hors-ligne
Avatar de l’utilisateur

3emeType

Cyanide Team

  • Messages: 122
  • Inscrit le: 23 Fév 2009, 16:34

Re: [AI] Bug Hillstat

Message21 Juin 2017, 12:58

suigfdhuibsyret a écrit:but its pretty hard if the ai doesnt even follow amador of all people on 85 effort on extreme.


I agree. That's why it is THE issue for me.

suigfdhuibsyret a écrit:I'm also pretty sure gilbert couldn't attack on the final hill, u see him standing up on his bike, meaning he's putting an immense effort.


He is standing only a few seconds and that does not imply he's putting an immense effort.

suigfdhuibsyret a écrit:Also explains why he cant beat dayer in the sprint with way way better secondary stats.


They have the same sprint stat: 70.
Hors-ligne

suigfdhuibsyret

  • Messages: 11
  • Inscrit le: 05 Août 2013, 23:25

Re: [AI] Bug Hillstat

Message21 Juin 2017, 13:20

3emeType a écrit:
suigfdhuibsyret a écrit:but its pretty hard if the ai doesnt even follow amador of all people on 85 effort on extreme.


I agree. That's why it is THE issue for me.

suigfdhuibsyret a écrit:I'm also pretty sure gilbert couldn't attack on the final hill, u see him standing up on his bike, meaning he's putting an immense effort.


He is standing only a few seconds and that does not imply he's putting an immense effort.

suigfdhuibsyret a écrit:Also explains why he cant beat dayer in the sprint with way way better secondary stats.


They have the same sprint stat: 70.


aight, i give up. This makes me way more angry than it should
Hors-ligne
Avatar de l’utilisateur

3emeType

Cyanide Team

  • Messages: 122
  • Inscrit le: 23 Fév 2009, 16:34

Re: [AI] Bug Hillstat

Message21 Juin 2017, 13:30

I've tried to escape with my train like you do but it doesn't work :(
Hors-ligne

Paul23

  • Messages: 10
  • Inscrit le: 24 Déc 2015, 22:32

Re: [AI] Bug Hillstat

Message21 Juin 2017, 14:33

Are you serious? His opponents were already "dead" before he started sprinting. This is ridiculous.
Hors-ligne
Avatar de l’utilisateur

3emeType

Cyanide Team

  • Messages: 122
  • Inscrit le: 23 Fév 2009, 16:34

Re: [AI] Bug Hillstat

Message21 Juin 2017, 14:56

Paul23 a écrit:Are you serious? His opponents were already "dead" before he started sprinting. This is ridiculous.


Are you talking about Stybar and Gilbert on Plouay?
If so, how can you say they are "dead"? They do the same sprint as Dayer, that means they are full with red energy before the sprint.
Hors-ligne

Paul23

  • Messages: 10
  • Inscrit le: 24 Déc 2015, 22:32

Re: [AI] Bug Hillstat

Message21 Juin 2017, 15:23

No, I'm talking about the LBL video, where you also said that his sprint stat mattered.

And in Plouay, he didn't even have a better SP stat than the others....

This game is so horribly broken, that it's not even funny. Beta testers have to defend themselves everywhere, because they get accused to be at fault for this mess.
Not trying to be mean, but you made a cycling game, which you do since what, 12 years?
And then it has the following features:

-AI is broken
-Sprints are broken
-hills are broken
-TTs are kinda broken
-the game is horribly optimized

And that's adding to the other bugs you listed already in another thread.
I bought your game since 2010. Every single time. And it feels like you take a step forward and 2 steps backward with every game. You cannot say, that it's not broken. It's that much broken, that communities are about to create current databases for PCM 15, because PCM 16 and 17 are broken.

I have
478h on PCM13
213h on PCM14
591h on PCM15
335h on PCM16

That's 1617h of gameplay within the last years. And I'm not counting PCM10, 11 and 12, since I don't have PCM10 on steam and the stats for PCM11 and 12 are bugged on my steam.
I enjoyed your games despite the flaws they have, but apparently you're not working well with the community.
And when you finally do, you mess it up. The AI was broken for several years. Nothing changed. Now you FINALLY change the AI and it's even more broken. And you argue, that it works as intended? You have got to be kidding.

You are so damn lucky, that there isn't a competitor on the market. You would sell nothing, if another company would make the same game.


Sorry for the rant. This is the last time I bought a PCM game and I will try to refund it.
Hors-ligne

Dalia

  • Messages: 79
  • Inscrit le: 08 Août 2012, 11:11

Re: [AI] Bug Hillstat

Message21 Juin 2017, 16:54

+1
Hors-ligne

trekbmc

  • Messages: 4
  • Inscrit le: 10 Déc 2014, 21:37

Re: [AI] Bug Hillstat

Message21 Juin 2017, 17:18

In my opinion, the main problem is that hill has now become a 'hill sprint' stat, it is a stat for hard uphill efforts over a short amount of time and it actually is a very good idea to have a stat like this, some riders (Ullisi springs to mind) can sprint very fast on a hill at the end of a flatish stage. The problem is that the mountain stat cannot be used for the short hills throughout the race before that - fleche wallonne is nothing like a multi mountain stage, even if they both finished on the Mur Du Huy!

The main difference is that puncheurs need to be able to main short, sharp efforts, no more than 1 or 2 minutes and sometimes even less, before descending, recovering and making another effort up the next hill, as well as being able to react when the peloton surges and bunches up with attacks and corners. Climbers on the other hand, need to be able to make an effort which might be up to an hour long or even more, they don't get much of a chance to recover as the effort is very consistent (aside from the occasional surge or attack and the sprint at the end, but your hill stat covers that now). The problem is that they are so different that they cannot be treated as the same thing just because they are both uphill. And the current method leads to a pure climber being stronger and getting less fatigued by the first 200km of Fleche Wallone than a pure puncheur.

I have a couple suggestions on the best way to fix this in general: The first is to simply embrace the fact that you have a 'hill sprint' stat and a 'climbing' stat and to adopt a third 'short term recovery' stat, which would be based on how quickly riders can expel lactic acid and be ready for another effort (how quickly they regenerate their yellow and red bars) this would relate to recovering both from surges from the bunch on the flat and more importantly after hills, so that they can be ready to sprint up hills again and again and again, which is exactly what a puncheur needs to do to win those sorts of races. Pure climbers however would have a low number in this stat and despite losing less bars on the hills when they are being ridden up at a pace lower than 85, they will be unable to make consecutive efforts up the short climbs and struggle towards the end of the race once they have been drained by all the repetitive efforts. This would be my preferred idea as it would increase the realism of how hilly races work in general (actually all types of races, some riders recover from efforts much better than others), however it may require the AI to be more realistic in how they race hilly races with more surges and bunching up like in real life.

A secondary solution is what Paul mentioned earlier with setting a default mo/hill ratio for the stage like in PCM14, but then being able to specify different ratios for different sections using the spline function in the stage editor, so I could rate a consistent mountain between kilometres 87.3-104.8 to have a ratio of 1, while a short climb between kilometres 117.2-122.4 might have a ratio of 0.5 and the remainder of the stage might have a ratio of 0 (full hill) (There's a better explanation here: http://pcmdaily.com/forum/viewthread.ph ... st_1241338), even with the method you could still keep the usage of the hill stat above 85 and for attacks so that there is some sort of differentiation between diesel climbers and punchy climbers and for the final sprint on a mountain.

Overall I am glad you guys are thinking up new ways to improve the game and listening to community feedback so thank you for that, just make sure improvements are better tested because otherwise they can be implemented with some flaws like the mountain stat being the most important stat for hilly stages like here.
PrécédentSuivant

Retour vers Pro Cycling Manager 2017

Qui est en ligne ?

Utilisateur(s) parcourant actuellement ce forum : Aucun utilisateur inscrit et 3 invité(s)

cron