In Season Development

henry296 · 04-12-2017, 12:46 PM

I'm trying to figure out if I can create more in-season roster shuffling outside of injuries. One type of move is to bring up a top prospect in June to avoid Super 2. Other than having that be a factor in the AI's logic, the other way to achieve that result is to have a players core ratings change throughout the season and have the scout's opinion change.

I've set scouting frequency to monthly but I'm wondering if the underlying ratings change or is it just running the scouting algorithm and the accuracy changes because of the experience of the scout.

The other option, is to make the AI looking more at stats than ratings to get more roster movement.

Any general thoughts on how to get the AI to bring more players up and down from the minors throughout the season. Most teams will use over 40 players in a season without counting, in OOTP it is probably in the low to mid 30s.

'94 EXPOS · 04-12-2017, 03:47 PM

Ratings do evolve throughout the year

Perhaps you are seeing the lower OOTP usage totals because of your injury settings?....the default setting is lower than the realistic one.

IMO a very high percentage of 25 man roster moves are health related. The exceptions: as you say, certain prospects will be strategically added at certain points during the season and there are also some guys who play themselves right off a roster. But this would account on average for one to three moves per team (again based on my experience).

henry296 · 04-12-2017, 05:28 PM

Quote:

Originally Posted by '94 EXPOS

Ratings do evolve throughout the year

Perhaps you are seeing the lower OOTP usage totals because of your injury settings?....the default setting is lower than the realistic one.

IMO a very high percentage of 25 man roster moves are health related. The exceptions: as you say, certain prospects will be strategically added at certain points during the season and there are also some guys who play themselves right off a roster. But this would account on average for one to three moves per team (again based on my experience).

I play with the highest injury settings.

Here is just a sample of 10 days in June for the Pirates last year. Some is injury related as well as brining up Taillon for one start and then sending him back down. Rob Scahill was in the minors for just a week and then brought back up with no obvious pitching injury.

June 11, 2016 Placed C Francisco Cervelli on the 15-day DL. Designated RHP Trey Haley for assignment. Recalled RHP Rob Scahill from Indianapolis (IL). Optioned LHP Kyle Lobstein and INF Cole Figueroa to Indianapolis. Reinstated RHP Arquimedes Caminero from the 15-day DL.
June 9, 2016 Optioned RHP Jameson Taillon to Indianapolis (IL). Recalled LHP Kyle Lobstein from Indianapolis.
June 8, 2016 Optioned RHP Curtis Partch to Indianapolis (IL). Recalled RHP Jameson Taillon from Indianapolis.
June 4, 2016 Optioned RHPs Rob Scahill and Wilfredo Boscan from Indianapolis (IL). Selected the contract of RHP Curtis Partch from Indianapolis. Recalled LHP Cory Luebke from Indianapolis. Sent RHP Arquimedes Caminero to Indianapolis for a rehab assignment. Transferred RHP Ryan Vogelsong to the 60-day DL.

NoOne · 04-12-2017, 05:43 PM

you'd need info from all teams over many years to understand the impact you had from the settings changes -- and assumes you have a baseline to compare from "default" in game as well as some RL-based target.

the one-up setting from "normal" is realistic rates of injuies.

player evaluation can be changed to use more stats... if your goal is more shuffling, i guess you could try tweaking that a bit, too. more stats = more knee-jerk reaction to small sample sizes. so, for your goal, i'd guess you want to use more stats and across less time - whether you use current or last year, just make 1 /3 years significantly more prevalent.

(argument for last year: during initial 'time' of a current year, those stats are not fully weighted until the sample size is 'larger' to warrant it -> per ootp feedback. so, any weight you put into current year gets dumped back into the other 2 in i assume a proportionate manner to their %'s in the settings. all that said.. i'd still suggest the current year being biggest weight relative to what you want)

henry296 · 04-13-2017, 10:59 AM

I experimented last night and I think I've found some settings that get me closer to my expectations. First, there is an option to base overall ratings in AI evaluation instead of Ratings. 2nd I reduced the weight on ratings to between 10-20 and then made current year stats between 60-70.

I think the next step is to increase the injuries to Very High from high and make some tweaks to increase DTD injuries and reduce longer term ones.

04-12-2017, 12:46 PM	#1
henry296 All Star Reserve Join Date: Mar 2002 Location: Baltimore, MD Posts: 514	In Season Development I'm trying to figure out if I can create more in-season roster shuffling outside of injuries. One type of move is to bring up a top prospect in June to avoid Super 2. Other than having that be a factor in the AI's logic, the other way to achieve that result is to have a players core ratings change throughout the season and have the scout's opinion change. I've set scouting frequency to monthly but I'm wondering if the underlying ratings change or is it just running the scouting algorithm and the accuracy changes because of the experience of the scout. The other option, is to make the AI looking more at stats than ratings to get more roster movement. Any general thoughts on how to get the AI to bring more players up and down from the minors throughout the season. Most teams will use over 40 players in a season without counting, in OOTP it is probably in the low to mid 30s.

04-12-2017, 05:43 PM	#4
NoOne Hall Of Famer Join Date: Apr 2015 Posts: 7,258 Infractions: 1/0 (0)	you'd need info from all teams over many years to understand the impact you had from the settings changes -- and assumes you have a baseline to compare from "default" in game as well as some RL-based target. the one-up setting from "normal" is realistic rates of injuies. player evaluation can be changed to use more stats... if your goal is more shuffling, i guess you could try tweaking that a bit, too. more stats = more knee-jerk reaction to small sample sizes. so, for your goal, i'd guess you want to use more stats and across less time - whether you use current or last year, just make 1 /3 years significantly more prevalent. (argument for last year: during initial 'time' of a current year, those stats are not fully weighted until the sample size is 'larger' to warrant it -> per ootp feedback. so, any weight you put into current year gets dumped back into the other 2 in i assume a proportionate manner to their %'s in the settings. all that said.. i'd still suggest the current year being biggest weight relative to what you want) Last edited by NoOne; 04-12-2017 at 05:45 PM.

04-12-2017, 03:47 PM	#2
'94 EXPOS Major Leagues Join Date: Jun 2011 Posts: 312	Ratings do evolve throughout the year Perhaps you are seeing the lower OOTP usage totals because of your injury settings?....the default setting is lower than the realistic one. IMO a very high percentage of 25 man roster moves are health related. The exceptions: as you say, certain prospects will be strategically added at certain points during the season and there are also some guys who play themselves right off a roster. But this would account on average for one to three moves per team (again based on my experience).

04-13-2017, 10:59 AM	#5
henry296 All Star Reserve Join Date: Mar 2002 Location: Baltimore, MD Posts: 514	I experimented last night and I think I've found some settings that get me closer to my expectations. First, there is an option to base overall ratings in AI evaluation instead of Ratings. 2nd I reduced the weight on ratings to between 10-20 and then made current year stats between 60-70. I think the next step is to increase the injuries to Very High from high and make some tweaks to increase DTD injuries and reduce longer term ones.