O.S.F.M. 2015 Recruiting(joint ops with Hybrid Set)

[WaitTilNextYear]

Re: O.S.F.M. 2015 Recruiting(joint ops with Hybrid Set)

Knight's charts are supposed to based off of the default ratings though, its just that they haven't been updated to reflect any changes to the ratings formulas in a long time. That's what I'm trying to do - find out what the exact changes are and update the charts to better reflect the default MLB ratings.

teeds even mentioned that a global edit was necessary to bring up the offense - that's alone says that Knight's charts need to be updated. Just taking a cursory look at the chart vs the default ratings, I can see contact, power, vision, and discipline potentially shifting to the right 1 standard deviation. Using Castro as an example, He has a 42/39 power rating which would equate to 11-12 HR's on Knight's chart, when his true weighted numbers put him in the 9-10 range (one standard deviation below). Same thing happens with Cano (default ratings) and numerous others across all hitting ratings, so there is some clout to my hypothesis. I just need to analyze the data and fit a line, and if it isn't that strong then a hotfix (like moving every category 1 to the right) will suffice.

Sure, I'm all for making improvements to ratings. We've already completely re-done speed and defense using a statistical base for SCEA players in what will be the 2015 Hybrid. Let's see what you come up with--as long as it works with the sim engine, I would be happy to adopt it and have it included. It would definitely need to be tested in sims as well.

The global edit that we did for the Hybrid roster last year was a +5 to plate discipline because the walks were too low. It might not be appropriate to assume this same flaw in the stock OSFM roster because the stock OSFM doesn't re-rate SCEA players based on ZiPS projections like we do/did.

Last year, in essence, we matched up batting and pitching ZiPS projections to Knight's spreadsheets, tested the data in a bunch of sims, and found that a +5 boost to disc was needed. From that process, it's hard for me to say for sure if the issue is the spreadsheets themselves being a bit off/dated or if it's only an issue when using ZiPS data as the input for the spreadsheet. I don't think the fact that we needed the fudge factor for what we did implicates any flaws in the spreadsheets for certain. Using a different projection system may work just fine with those spreadsheets. We have not isolated the variable yet, although it could be that the spreadsheet scales have become 'out of phase' with the sim engine as you're (I think) suggesting.

[Willard76]

Re: O.S.F.M. 2015 Recruiting(joint ops with Hybrid Set)

Knight's charts are supposed to based off of the default ratings though, its just that they haven't been updated to reflect any changes to the ratings formulas in a long time. That's what I'm trying to do - find out what the exact changes are and update the charts to better reflect the default MLB ratings.



teeds even mentioned that a global edit was necessary to bring up the offense - that's alone says that Knight's charts need to be updated. Just taking a cursory look at the chart vs the default ratings, I can see contact, power, vision, and discipline potentially shifting to the right 1 standard deviation. Using Castro as an example, He has a 42/39 power rating which would equate to 11-12 HR's on Knight's chart, when his true weighted numbers put him in the 9-10 range (one standard deviation below). Same thing happens with Cano (default ratings) and numerous others across all hitting ratings, so there is some clout to my hypothesis. I just need to analyze the data and fit a line, and if it isn't that strong then a hotfix (like moving every category 1 to the right) will suffice.

That info will be invaluable once you key in on it. Might bring our hybrid ratings up a bit at the same time and all we need to do once you find out the difference would be to but a formula into the cells on our spreadsheet to give us the new numbers.

[zbundy3]

Re: O.S.F.M. 2015 Recruiting(joint ops with Hybrid Set)

Thought this was pretty cool. Tomas being rendered today for the game, not sure if it's for this year or next

Looks like the Pirates are getting their faces scanned today, here is sean rodriguez getting his done. This was from the pirates instagram account. https://instagram.com/pittsburghpirates/

[The Kid 24]

Re: O.S.F.M. 2015 Recruiting(joint ops with Hybrid Set)

Really hoping this is for 15... Could face scans be the final step for rosters?

[Willard76]

Re: O.S.F.M. 2015 Recruiting(joint ops with Hybrid Set)

No. They do this every year for the next

[ShowTyme15]

Re: O.S.F.M. 2015 Recruiting(joint ops with Hybrid Set)

Really hoping this is for 15... Could face scans be the final step for rosters?

To piggy back on what Willard said. These facial scans are for MLB 16. SCEA does facial scans in spring training each year.

Example facial scans for MLB 14 were done in Spring Training 2013. MLB 15 scans were done in spring of 2014.

[YankeeNO2]

Re: O.S.F.M. 2015 Recruiting(joint ops with Hybrid Set)

I am doubtful, but I can promise the one we have created for our set thanks to Motown looks as good if not better than the real thing

Looks awesome!

[bacon96]

Re: O.S.F.M. 2015 Recruiting(joint ops with Hybrid Set)

Someone will have to take the Brewers this year... Brad and myself won't be able to work on them, can't find enough time to commit to do a good job.

If no one else does the Brewers then I'd like to take a stab at it. Let me know.

[WaitTilNextYear]

Re: O.S.F.M. 2015 Recruiting(joint ops with Hybrid Set)

If no one else does the Brewers then I'd like to take a stab at it. Let me know.

No one's gonna argue with bacon. Go for it.

[bacon96]

Re: O.S.F.M. 2015 Recruiting(joint ops with Hybrid Set)

No one's gonna argue with bacon. Go for it.

Thanks man.

Who do I talk to about this then?

[WaitTilNextYear]

Re: O.S.F.M. 2015 Recruiting(joint ops with Hybrid Set)

Thanks man.

Who do I talk to about this then?

Just start a new thread for your team where you can get a roster list/ratings/potentials together etc.. That thread is also good for recruiting other fans of your team to help so you don't have to do all the faces/ratings/research/equipment...I've never OSFM'd a team before, but you can ask some of the other guys around what it is exactly that they do and how they do it. You can also drop RidinDwnKingsley a PM as well to let him know your desire to contribute. He's curating the OSFM effort this year in Knight's place.

[rjackson]

Re: O.S.F.M. 2015 Recruiting(joint ops with Hybrid Set)

Sure, I'm all for making improvements to ratings. We've already completely re-done speed and defense using a statistical base for SCEA players in what will be the 2015 Hybrid. Let's see what you come up with--as long as it works with the sim engine, I would be happy to adopt it and have it included. It would definitely need to be tested in sims as well.

The global edit that we did for the Hybrid roster last year was a +5 to plate discipline because the walks were too low. It might not be appropriate to assume this same flaw in the stock OSFM roster because the stock OSFM doesn't re-rate SCEA players based on ZiPS projections like we do/did.

Last year, in essence, we matched up batting and pitching ZiPS projections to Knight's spreadsheets, tested the data in a bunch of sims, and found that a +5 boost to disc was needed. From that process, it's hard for me to say for sure if the issue is the spreadsheets themselves being a bit off/dated or if it's only an issue when using ZiPS data as the input for the spreadsheet. I don't think the fact that we needed the fudge factor for what we did implicates any flaws in the spreadsheets for certain. Using a different projection system may work just fine with those spreadsheets. We have not isolated the variable yet, although it could be that the spreadsheet scales have become 'out of phase' with the sim engine as you're (I think) suggesting.

My two points of constructive criticism of the Hybrid are speed/defense (which it appears you may have addressed) and that the linear scale was too low. On the other end of the spectrum, SBFM was too high. I thought that it might have been obvious, though, since we had to edit all of the draft classes each season since they would have a number of them that came in as better already than established MLB regulars. I was planning on just using a base OSFM v2 this year due to those issues as then I'd only have to edit potentials for those teams that went too homer with their team and that would only have to be done once.


That said, I wish that I had the time to contribute to this more (like you guys do and I'm appreciative), so this is the best I can do - provide constructive criticism whenever I can.

[teeds]

Re: O.S.F.M. 2015 Recruiting(joint ops with Hybrid Set)

My two points of constructive criticism of the Hybrid are speed/defense (which it appears you may have addressed) and that the linear scale was too low. On the other end of the spectrum, SBFM was too high. I thought that it might have been obvious, though, since we had to edit all of the draft classes each season since they would have a number of them that came in as better already than established MLB regulars. I was planning on just using a base OSFM v2 this year due to those issues as then I'd only have to edit potentials for those teams that went too homer with their team and that would only have to be done once.


That said, I wish that I had the time to contribute to this more (like you guys do and I'm appreciative), so this is the best I can do - provide constructive criticism whenever I can.

We appreciate the feedback and welcome it but could you clarify what you mean by our "linear" scale? We don't have our own scale and use knights charts to get our numbers. Is it the look of the ratings that seem low or did your sim results reveal them to be too low? We ran countless sims to get out numbers in line with league averages and came pretty damn close.

One point of contention that we who work on the project had were our speed and fielding ratings. WTNY and I talked about this a few weeks back and he developed new formulas using defensive metrics that have now been implemented into our ratings. These are reflected on our spreadsheet and we're excited by the results. Another thing we started with our last release this past summer was batter splits and we are going to take a further look at that.

[Willard76]

Re: O.S.F.M. 2015 Recruiting(joint ops with Hybrid Set)

Is anyone interested in jumping on creating llamas to run around during Cactus League games?

[Willard76]

Re: O.S.F.M. 2015 Recruiting(joint ops with Hybrid Set)

Sorry the person responsible for my previous post has been sacked