Inside CH2K8: Research and Experiments

[Taf1971]

Re: Inside CH2K8: Research and Experiments

What I was saying is If we input the stats from ESPN into a spreadsheet and based on certain %'s shooting etc, how many steals per game would be considered an A "98" rating etc. blocks etc. that it would give us the rating by position as much as possible. It would also be cool to "rerate" the alumni rosters too based on this criteria. If you have a roster spreadsheet that pulls the stats into a rating that would be awesome to share. It could be a good starting point for other ideas.

[vgunn]

Re: Inside CH2K8: Research and Experiments

Using Sim With Hoopcast to simulate games where one team is at least 10 points better in Team Overall rating, the higher rated team was 100-10. This ratio tended to hold for most of the sample, and was closer a 7-1 split about midway through.


This means that, using Hoopcast to simulate, the clearly superior team would win no more than 10 out of 11 games. These are non-conference matchups, several of which should be blowouts. FWIW, most of the upsets I saw were small schools beating mid-majors.

So I assume that simming with Hoopcast is much more accurate than fast simming a game. Correct?

[OSUPiper]

Re: Inside CH2K8: Research and Experiments

So I assume that simming with Hoopcast is much more accurate than fast simming a game. Correct?

I can't tell exactly, but I think it is. I may run some tests at some point.


What I like about Hoopcast is this: I "re-simulate" upsets in legacy, and Hoopcast allows me to easily track games I have marked in case I need to re-sim without having to go look up the final score from another screen.

[GAMEBREAKER85]

Re: Inside CH2K8: Research and Experiments

Did you ever come up with a good template for player tendencies?

I did a fake roster and re-did the tendencies but the issue is incoming recruits in a legacy have very low shot tendencies compared to what I used for my edits.

[Brock]

Re: Inside CH2K8: Research and Experiments

Did you ever come up with a good template for player tendencies?

I did a fake roster and re-did the tendencies but the issue is incoming recruits in a legacy have very low shot tendencies compared to what I used for my edits.

You could add 25+ to the tendencies. I usually do a mixture of that and what the players strong skills are. Can't shoot, but have high layups/dunk just boost the drive tendencies, etc.