Simulator Design

[Fire]

I have a big task for someone if they want to take it. I need someone to come up with a very detailed design for a simulator.

Given 2 teams I need a design to determine the result of the game. The result must be determined by ratings of the players(offense, defense, physical, intangible, etc?) and goalies (overall rating) involved. Of course some randomness has to be involved in order for weaker teams to occasionally beat stronger teams.

The info I need from this simulation per game is:

- The final score
- The goal and assist scorers (need to be based on the lines decided by the GMs)
- Penatlies taken by whom and when
- Time of the goals
- # number of shots
- PP and PK stats
- Goalie pulled? and when?
- 3 Stars
- etc?

Of course the simulator can be more or less detailed than this. It depends on what level of detail the GMs want out of the Simulator.

If someone can come up with a detail design on how to do this I can do the programming part. If we can get a simulator integrated with the rest of our league the possibilities for improvement for the league is great.

If anyone want to take on this task let me know and we can work on together.

[JohnnyFlame]

I would ask for a cutback on the number of penalities especially the ridiculous number of misconducts for kicking some guy in the head for instance. With that just in general make it harder to score and that with some randomness should give the lower teams a better go of it and make it far more realistic.

[Cheese]

We need SOMEONE to help "constructively" with this Simulator. Random comments dont cut it. IF anyone has the time to help Cam put it together it will benefit us all in the longrun, as we can dump the FHL Simulator.
It is not onerous...simply a task that has you using your grey matter with thoughts to what we want to see here.

[JiriHrdina]

Cheeze, perhaps you or Cam can provide more detail regarding what this entailes? When I first read the post I assumed I didn't have the "know-how" for something like this as I've never done a sim design, however is that experience necessary? I guess I'm just not sure what this specifically entails and if its something I would have the skills to do.

[Cheese]

sure thing Jiri!
This isnt a programming thang...simply a process in how we want the sim to work, as Cam laid out in his post. I was going to help but my time is limited and perhaps my brain is too. When a complete scope is laid out I will scan it and add my 5 cents.
Cam can help you with how he wants the info and in what format...but its not onerous...simply a commitment to start and FINISH it by years end or before...so we can beta test it.
Of course anyone with programming experience would be welcome to assist...but formulating the plan is key...Cam is a genious at putting ideas into action. ;)
Email Cam or put your name forth here.....the more the merrier as long as its coordinated.

[kermitology]

I've got an idea for the PK and PP stuff.. I don't know what kind of settings you can play with.. but here goes..

Take a league average for both PK and PP, from the NHL, and start with that for each team. Then.. maybe a bit of randomness, sprinkle the CPHL with some + or - a certain level for each team, just so everyone is different. But maybe take away from one part and put it in the other. So if you make a team have less than league average PP boost the PK percentage.

Just a thought..

[Cheese]

Originally posted by kermitology@Aug 4 2004, 12:45 PM
I've got an idea for the PK and PP stuff.. I don't know what kind of settings you can play with.. but here goes..

Take a league average for both PK and PP, from the NHL, and start with that for each team. Then.. maybe a bit of randomness, sprinkle the CPHL with some + or - a certain level for each team, just so everyone is different. But maybe take away from one part and put it in the other. So if you make a team have less than league average PP boost the PK percentage.

Just a thought..
again guys...random comments dont cut it...volunteer or not...email Cam and step up to the plate.

[Hanna Sniper]

If anyone comes across anything readable about “building a sim” or any “how to” please post this way maybe we can see what’s involved and whats needed.

Otherwise I’m clueless (yea, yea, keep it to yourself)

[Fire]

Maybe we should start out more simpler and just have the sim spit out the score of game and provide no stats. We can grow to more complicated stuff later.

Ok, say you have the following information at your disposal to sim a game:

- 2 CPHL teams
- Player Ratings (Offense(0-100), Defense(0-100), Physical(0-100), Intangibles(0-100))
- Goalie Ratings (Just an overall rating for them)
- Line Combinations

Given the above, you need to generate the result of the game.

Tell me how you think this should be done. All I want for now is the score.

Any questions?

[Beerfund]

Cam you have a PM. I think.

Edit: Ooops never mind - not only do you know this, you've responded :blink:

[moon]

If there is no computer knowledge needed and limited math skills then I will gladly help out.

[Fire]

Originally posted by moon@Aug 5 2004, 02:22 AM
If there is no computer knowledge needed and limited math skills then I will gladly help out.
Computer knowledge... no.

Math Skills... yes, somewhat.

[MJK]

I would be more than happy to help out anyway that I can. I have a ton of time on my hands!

[Fire]

Originally posted by mjk123@Aug 10 2004, 09:31 PM
I would be more than happy to help out anyway that I can. I have a ton of time on my hands!
I think the best way is for everyone who is interested to come up with a basic outline on how you think it should work. Post it here if you want.

[Shawnski]

No question, this is a very large task, IMO.

From a business analysis perspective, two questions come to mind.

1) Why change?

2) If a change is necessary, what is good about the current system and what is not good.

Cam, I can assist. Have both comp skills and math skills, as well as logic development background in large systems. This is basically what I do for a living...... so.......where do I send the bill? :)

[Fire]

1) Why change?

It will be a lot less work on our part because the simulator can be integrated into the rest of the cgis. Thus more automation can be done.

2) If a change is necessary, what is good about the current system and what is not good.

I don't think the current simulator is terrible on how it determine games (though how they do it is unknown). If we could make a simulator that generates similar output to the current simulator I think that would be great.


"This is basically what I do for a living...... so.......where do I send the bill?"

I've been wondering that too. :)

[Sylvanfan]

Send the Bill to Cheese, he's got lot's of money!!