OSUPiper's CH2K8 CPUvCPU Experience

[OSUPiper]

Re: OSUPiper's CH2K8 CPUvCPU Experience

Thanks. 2 other questions that have come up as I'm getting back into the game after a few years.

1) The coaching setting for Offensive Tempo... is this kind of a modifier on overall tendencies? Like the probability for a specific player to shoot at any given time is a function of (player tendency) * (tendency slider) * (offensive tempo). That's vaguely how I see it, but I'm interested in your take.

2) Have you ever tried using custom playbooks, i.e. from the esoteric information thread?

1. As I understand it, it is the pace at which a team moves the ball up the floor and score in transition.


2. No, but I could see an application if I had time to get deep enough into it.

[doctorhay53]

Re: OSUPiper's CH2K8 CPUvCPU Experience

been playing with the set on page 21, the latest set posted. For my tastes, guards don't drive the ball enough; outside players who can shoot are far more valuable than guys with some handles who can drive.

Based on my reading of the thread, V4.0 would be the set with the most drives?

edit: I guess what's missing for me is a guard driving from the 3pt line and ending with a traditional layup animation. Most drives end in a foul or some other animation. I think the only way to do it would require sacrificing a lot of defense. So it's probably tough to do without sacrificing alot statistically.

[Gtl420]

Re: OSUPiper's CH2K8 CPUvCPU Experience

Just so y’all know. I told OSUPiper this earlier in the thread but these sliders also play well for human vs cpu. I think the only thing I changed was the shooting sliders and made them higher


Sent from my iPhone using Operation Sports

[doctorhay53]

Re: OSUPiper's CH2K8 CPUvCPU Experience

I really prefer custom playbooks because the dynamic of play is just so much better. Real driving layups and pick and roll dunks and everything. With "All Plays" the stats are decent but I just find the flow of the game to be ugly.

However, custom playbooks then seem to mess up the stats a bit with the latest sliders. So I'm a bit torn on what to do.

As far as custom playbooks go, several years ago I sat down in practice mode and tested every single play in the game. Some plays, for instance, simply stop working after the first pass nearly all of the time. Against man and 2-3 zone, I wrote down who got a good shot and only kept plays that flowed pretty well.

For a given custom playbook, I put in a number of plays for a given position; for instance, the best scorer gets 6 plays, next best gets 5, etc. For 5 starters this gets us to 20 plays. I have a python script to randomly generate a playbook.

I've found that using a custom playbook vs "all plays" results in the custom playbook usually winning by a ton. In CH2k8 there is definitely an effect where good offense turns into good defense.

If I use the same custom playbook with both teams, I've gotten some of the best games to watch that I've ever seen. Really close games, home teams with bigger comebacks in the last 5 minutes, etc. But again, it messes up the stats.

OSUPiper, if you ever come back, I'd love to give you a custom playbook and just have you run it through your system for a little while. I just want to know how "off" the stats are, because I really want to watch the game with custom playbooks.

[Brock]

Re: OSUPiper's CH2K8 CPUvCPU Experience

I really prefer custom playbooks because the dynamic of play is just so much better. Real driving layups and pick and roll dunks and everything. With "All Plays" the stats are decent but I just find the flow of the game to be ugly.

However, custom playbooks then seem to mess up the stats a bit with the latest sliders. So I'm a bit torn on what to do.

As far as custom playbooks go, several years ago I sat down in practice mode and tested every single play in the game. Some plays, for instance, simply stop working after the first pass nearly all of the time. Against man and 2-3 zone, I wrote down who got a good shot and only kept plays that flowed pretty well.

For a given custom playbook, I put in a number of plays for a given position; for instance, the best scorer gets 6 plays, next best gets 5, etc. For 5 starters this gets us to 20 plays. I have a python script to randomly generate a playbook.

I've found that using a custom playbook vs "all plays" results in the custom playbook usually winning by a ton. In CH2k8 there is definitely an effect where good offense turns into good defense.

If I use the same custom playbook with both teams, I've gotten some of the best games to watch that I've ever seen. Really close games, home teams with bigger comebacks in the last 5 minutes, etc. But again, it messes up the stats.

OSUPiper, if you ever come back, I'd love to give you a custom playbook and just have you run it through your system for a little while. I just want to know how "off" the stats are, because I really want to watch the game with custom playbooks.

I'd be interested in the custom playbook/python script.

[doctorhay53]

Re: OSUPiper's CH2K8 CPUvCPU Experience

I'd be interested in the custom playbook/python script.

Here you go. Let me know if you need any help.

https://github.com/bhayden53/ch2k8pb

edit: I've noticed an issue where sometimes the while loop never breaks. I will need to undertake an actual cleaning of the code at some point. I whipped it up for myself over the weekend, converting an old script in a different language that I wrote in like 2010.

edit2: pushed a horrible hack to the infinite while loop issue. With the current settings of the config file the PG always has the same 2 plays selected. My PG is the weakest player on my team.

Interestingly though, one of the main reasons why I want to use custom playbooks is that PG's almost never shoot the ball in the flow of the offense. You really have to call a specific play that tends to end in a PG shot to get them to score in my experience.

[doctorhay53]

Re: OSUPiper's CH2K8 CPUvCPU Experience

I'm not totally happy with my categorization of some of the plays. I think the full playbook needs another review using the current slider settings (shot tendencies will greatly affect who shoots at what point in some of the plays).

[Brock]

Re: OSUPiper's CH2K8 CPUvCPU Experience

Thanks, I'll def check it out.

[doctorhay53]

Re: OSUPiper's CH2K8 CPUvCPU Experience

Thanks, I'll def check it out.

I just pushed some changes. The readme hasn't been updated but I did actually go through every play. Since my goal is to be able to get specific players to score, I simplified the categories of each play to basically only cover plays where the player actually tends to try to score.

Another thing is that in the config file, for each position, the "skills" list has an entry for each play I want. i.e. my SG is a heavy 3pt shooter so in the config file I used ["3pt","3pt","3pt","3pt","inside","iso"] so that 4 of the plays selected for the SG will be 3pt plays. I had noticed that sometimes the playbooks were a little too random in terms of what types of plays a player would get.

I'm still torn on using custom playbooks. I find it does make the game a bit more challenging in cpu vs cpu (road games are really dicey), but I also haven't 100% settled on a method for creating a playbook that creates the types of shots I'd like to see.

[kube]

Re: OSUPiper's CH2K8 CPUvCPU Experience

Can someone please tell me how to find the spot for shot sick ft off ? I do not want to shoot free throws with the shot stick but I do not know how to do that.