Stars versus 20 /80

[Padreman]

I've been playing this game since version 13 and I've always use the stars for player potential and current ability to me it just makes sense. However, I know sometimes the stars rating method can be very misleading. My question is which method is more superior? Which method best describes the players ability and potential? I'll give you example I was looking at a three star player yesterday then I switch the ratings method to 20/80 and he went down to a 44 which I guess will be a little bit below average maybe average . To me I Would have thought it would've been a higher rating like a lease in the 50s

[ThePretender]

They're both misleading and I'd avoid using them. You'll be better off ignoring it. Lots of times a 1 star or 20 overall guy can be a quality starter or a 4-5 star or 70+ player will suck.

[Padreman]

Originally Posted by ThePretender

They're both misleading and I'd avoid using them. You'll be better off ignoring it. Lots of times a 1 star or 20 overall guy can be a quality starter or a 4-5 star or 70+ player will suck.

Yeah that's finding out I have a five star pitcher they lost 22 games it can really screw you up

[NoOne]

Originally Posted by Padreman

I've been playing this game since version 13 and I've always use the stars for player potential and current ability to me it just makes sense. However, I know sometimes the stars rating method can be very misleading. My question is which method is more superior? Which method best describes the players ability and potential? I'll give you example I was looking at a three star player yesterday then I switch the ratings method to 20/80 and he went down to a 44 which I guess will be a little bit below average maybe average . To me I Would have thought it would've been a higher rating like a lease in the 50s

it's just the rounding when converting to another scale, and possibly even how scouting inaccuracy is applied?

even with 100% scouting accuracy there's a grey area around any particular star rating when converting to another scale, since the resolution is so minimal when it's 1-5.

It's not that any scale is superiour to another. it's more about personal preference.

e.g. the larger the scale and the more accurate the scouting then it gets quite easy to know for a fact who the good players are.

so, if you think it's too easy now you probably don't want a larger scale. if yyou think it's a bit too clandestine, then you probably want a larger scale... bump up one at a time until you like the feedback that you see in the ratings.

5 star is only for overall / potential correct? the individual ratings only go down to 1-10? This stuff applies moreso to the individual ratings than the overall / potential rating - the nature of a rating that is comprised of many other ratings.

overall / potential can be deceiving. as suggested above i'd reduce how much you look at overall / pot. focus more on the individual ratings and which traits = the role you are looking for.

you likely don't want an overall 75/80 with 22/80power in an rbi producing spot in the lineup when you can get a overall 50 with some power that can actually drive in runs at a higher rate (objective fact, not grey area for this example - can only prove through simulating and comparing results - assumign in this example this is true). his obp, eye isn't as important in this role --

you can come to this conclusion and adjust, but an overall rating never adjusts for the situation. it's for the whole pie -- the total average of all roles and situations that occur in a particular proportion. so, when context is enough different from "average" you need to change how you value that overall rating.

us it for filtering and not player evaluation... and since you are familar with its inaccuracy, make sure you go1-2 stars below what you actaully are looking for. typically that plus age is enough for me to make a large list manageable.

i don't quite share the opinion it's as extreme as a 20 being better than someone in the 70's (never seen it myself), but overall definitley does not tell you who is better very often just by comparing it, so don't use it that way unless they are the same type of player (2+ players wiht avg/obp with no power... or 2 different power hitters with similar traits, etc etc -- even then i'd still focus on the individual traits and go with what i like more.)

[Hrycaj]

I have been a stats only guy for more than a few versions now. I really became tired of trying to figure out the accuracy of stars and ratings.

Although, when I did play with stars isn't there a setting that determines if the stars are based on overall talent or positional ratings? If you are set to ratings relative to position it would explain why your 5 star 2nd basemen could hit .250 with zero power. He may be a 5 star 2nd basemen but that just proves that your 2nd base talent league wide is poor overall.

So under that premise I would think 20/80 is more accurate. However, I could easily be convinced otherwise!

[NoOne]

yeah, you can make them relative to position and/or league too.

the higher the resolution in the scale the better it is for you to understand ability. 1-100 will show you the most (20-80 limit only on overall - potential, which are mostly useless / ignored ratings anyway).

you probably didn't like ratings (or thought they weren't as predictive as stats) becasue you were using overall too much when comparing alternative choices. i can tell you for sure that normal scouting with 4-6M scouting budget for that categor plus a great scout is handsdown more predictive that stats.... if you use a 1-5 scale it may not be true (not even sure if individual ratings go below 1-10?)... i use larger scales. 1-10 i would guess is borderline if weak budgets, average scout. but, if you use 1-20 or greater, the individual ratings with the qualifications i gave a few sentences previous is better.

The 'relative' options for ratings adds another layer for you to decipher... some like it... i do not. i stick to one scale and i don't have to translate or convert things in my head. also, if relative to current talent, that makes it float year-to-year a small amount (up and down)... kis(s) principle. (keep it simple, stupid)

e.g. what adds up to a 80/80 or 5/5 in 2016 won't be the same in 2017 or 2158 - good luck figuring that out with the naked eye. you'd have to analyze distribution of ratings for each position to know what adds upt to "best" in a particular year or whenver it decides the scale - possibly continuous in nature.

if you've played with these option for years, it's likely second nature, but it's not as clear (objective fact), if you use "relative to position" no matter how comfortable you are. I'm assuming relative to league is not based on the players in the league but rather the league level setting... if i am wrong and it's based on current talent in the league, then this applies to that also.

[Hrycaj]

Originally Posted by NoOne

yeah, you can make them relative to position and/or league too.

the higher the resolution in the scale the better it is for you to understand ability. 1-100 will show you the most (20-80 limit only on overall - potential, which are mostly useless / ignored ratings anyway).

you probably didn't like ratings (or thought they weren't as predictive as stats) becasue you were using overall too much when comparing alternative choices. i can tell you for sure that normal scouting with 4-6M scouting budget for that categor plus a great scout is handsdown more predictive that stats.... if you use a 1-5 scale it may not be true (not even sure if individual ratings go below 1-10?)... i use larger scales. 1-10 i would guess is borderline if weak budgets, average scout. but, if you use 1-20 or greater, the individual ratings with the qualifications i gave a few sentences previous is better.

The 'relative' options for ratings adds another layer for you to decipher... some like it... i do not. i stick to one scale and i don't have to translate or convert things in my head. also, if relative to current talent, that makes it float year-to-year a small amount (up and down)... kis(s) principle. (keep it simple, stupid)

e.g. what adds up to a 80/80 or 5/5 in 2016 won't be the same in 2017 or 2158 - good luck figuring that out with the naked eye. you'd have to analyze distribution of ratings for each position to know what adds upt to "best" in a particular year or whenver it decides the scale - possibly continuous in nature.

if you've played with these option for years, it's likely second nature, but it's not as clear (objective fact), if you use "relative to position" no matter how comfortable you are. I'm assuming relative to league is not based on the players in the league but rather the league level setting... if i am wrong and it's based on current talent in the league, then this applies to that also.

Everything you say here makes a lot of sense. I never wanted to have anything too precise anyway. For me that takes the fun away. What I also learned about myself was that I would have a real hard time playing a lower rated/star guy over a higher one. I would keep thinking that the higher one has to be better at some point and in the end I was making poor roster decisions because of stars and ratings. I found it way more enjoyable to just turn all of that off. It was very liberating to simply play guys based on their performance without knowing what they were really rated. However, I'm not one of those stats only dudes that want to shove it down people's throat. My player screens are not that colorful and do look empty. Also, Padreman specifically asked about ratings so I won't do anything more to derail this discussion. I think you have pretty much answered the question as well as it could be.

[PSUColonel]

what I would like to see is a scouted grade (not the overall AI evaluation) that real MLB teams use. It is on a 20-80 scale, but like the individual tool grades, it's a system based on ½ grades....so you would never have a 57...it would to either be a 55 or 60.

[SxSnts9]

Originally Posted by PSUColonel

what I would like to see is a scouted grade (not the overall AI evaluation) that real MLB teams use. It is on a 20-80 scale, but like the individual tool grades, it's a system based on ½ grades....so you would never have a 57...it would to either be a 55 or 60.

I stated this exactly in the suggestions for OOTP 18 thread. If my ratings for things like contact, gap power, avoid K come in the intervals of 5 like MLB uses...why does overall/potential come in 1's?

[NoOne]

Originally Posted by Hrycaj

Everything you say here makes a lot of sense. I never wanted to have anything too precise anyway. For me that takes the fun away. What I also learned about myself was that I would have a real hard time playing a lower rated/star guy over a higher one. I would keep thinking that the higher one has to be better at some point and in the end I was making poor roster decisions because of stars and ratings. <Snipped, by reply>

Exactly.. I've never used a 5-pt scale... there are probably useful 1 star guys mainly due to rounding errors, so turning stars off would probably be best, lol...

when you use a larger scale it actually can be useful for cutting out 100's-1000's of players from a list. if i am starting a 20 or 21 / 80 player... i've done soemthign terribly wrong... doesn't even happen when i turn draft pick trading off

outside of starting players and key pitchers, def have some lowly rated palyers on a 25-man / 40-man rosters.

Edit: and if you do 20-25-30 etc... it's just a 1-to-13 scale... if they do that, add the 1-100 to overall, because there's aready a few small resolution scales.