NBA 2K13 Player Ratings - Kobe - 93, Melo - 92, Rondo - 90

[LD2k]

[youALREADYknow]

Quote: Originally Posted by LD2k

Real talk - on a sidenote - feel free to leave feedback to all ratings. Feedback is great for the 2K Insider my friends.

As an outside observer, it seems that might be the problem to begin with. I'm sure you have a better idea than any of us of the current process, but the results speak for themselves as far as 2K11 and 2K12 are concerned.

Objective and meaningful sports data is more accessible than at any time in history and most common advanced metrics are available for free. A few Python scripts and a licensed data source to crawl/scrape could generate a database full of enough data to programmatically come up with at least 75% of the attributes, tendencies, hot spots, and hot zones.

It's quite a shame to see the hard work of the designers and development team over there blemished by the inputs that go into the game. Arguing about overall ratings is largely trivial and full of speculation, but many forum members are justified in their concern given how individual attributes have been inflated in the past in response to "feedback", media-driven hype, or performance in small sample sizes.

Accepting the community's feedback on a case-by-case basis just flat out misses the point regardless of the sincerity of the approach. The only people who can give meaningful ratings advice on 2K13 are the ones who have access to the sim engine formulas and logic that produce statistical output.

[DukeC]

....

I think a couple of sig skills might be inflating the overall of some of these dudes. Especially if they have more than 1.

Won't be able to test this theory till I get the game in my hands though. Hopefully we'll be able to edit sig skills (Give them as well as take them away). Unless they are able to assign the sig skills in part due to certain ratings....

[JwP23]

Quote: Originally Posted by SteboSSK

You dont redeem yourself in a matter of a week or 2 of good ball. Everyone knows Melo can play and well at that, but based on the 2k system there is no way he could be a 92 w/o overrating alot of his abilities to boost the OVR. Alot of credible roster makers had Melo at around 88-89 and even 2k's last 2k12 update had him at 90. So unless the ratings have boosted scale wise no way his play this past season garners a 2 point bump. Even being a big Kobe fan I can admit some of his abilities particularly his defense is a tad overrated. Kobe should be around a 90-91 if KD is a 95.

^A very knowledgable Lakers fan who sees the truth.

[iLLosophy]

Quote: Originally Posted by youALREADYknow

As an outside observer, it seems that might be the problem to begin with. I'm sure you have a better idea than any of us of the current process, but the results speak for themselves as far as 2K11 and 2K12 are concerned. Objective and meaningful sports data is more accessible than at any time in history and most common advanced metrics are available for free. A few Python scripts and a licensed data source to crawl/scrape could generate a database full of enough data to programmatically come up with at least 75% of the attributes, tendencies, hot spots, and hot zones. It's quite a shame to see the hard work of the designers and development team over there blemished by the inputs that go into the game. Arguing about overall ratings is largely trivial and full of speculation, but many forum members are justified in their concern given how individual attributes have been inflated in the past in response to "feedback", media-driven hype, or performance in small sample sizes. Accepting the community's feedback on a case-by-case basis just flat out misses the point regardless of the sincerity of the approach. The only people who can give meaningful ratings advice on 2K13 are the ones who have access to the sim engine formulas and logic that produce statistical output.

Even with 3rd party data, there's many times where data isn't properly transformed to fit the NBA 2k logic. For example, in the other thread I was talking about how, even though Iggy shot great from 3pt land, he doesn't deserve a high 88-90 rating because IN GAME guys with that high of a rating can hit multiple contested 3's in any given game. Iguodala got looks that gave him a lot of space to shoot - so he could've shot the same % if he had an 80 three point rating - his great shooting had more to do with the offense they ran versus his deadeye shooting. Same could be said for most of the 3pt shooters on Orlando - they benefited from everybody being focused on Dwight, so they get A LOT of open looks from 3 which results in high percentages. That doesn't necessarily warrant an 88 three pointer for everyone.

[Colts18]

Quote: Originally Posted by youALREADYknow

As an outside observer, it seems that might be the problem to begin with. I'm sure you have a better idea than any of us of the current process, but the results speak for themselves as far as 2K11 and 2K12 are concerned. Objective and meaningful sports data is more accessible than at any time in history and most common advanced metrics are available for free. A few Python scripts and a licensed data source to crawl/scrape could generate a database full of enough data to programmatically come up with at least 75% of the attributes, tendencies, hot spots, and hot zones. It's quite a shame to see the hard work of the designers and development team over there blemished by the inputs that go into the game. Arguing about overall ratings is largely trivial and full of speculation, but many forum members are justified in their concern given how individual attributes have been inflated in the past in response to "feedback", media-driven hype, or performance in small sample sizes. Accepting the community's feedback on a case-by-case basis just flat out misses the point regardless of the sincerity of the approach. The only people who can give meaningful ratings advice on 2K13 are the ones who have access to the sim engine formulas and logic that produce statistical output.

Preach man. Preach.

[youALREADYknow]

Quote: Originally Posted by iLLosophy

Even with 3rd party data, there's many times where data isn't properly transformed to fit the NBA 2k logic. For example, in the other thread I was talking about how, even though Iggy shot great from 3pt land, he doesn't deserve a high 88-90 rating because IN GAME guys with that high of a rating can hit multiple contested 3's in any given game. Iguodala got looks that gave him a lot of space to shoot - so he could've shot the same % if he had an 80 three point rating - his great shooting had more to do with the offense they ran versus his deadeye shooting. Same could be said for most of the 3pt shooters on Orlando - they benefited from everybody being focused on Dwight, so they get A LOT of open looks from 3 which results in high percentages. That doesn't necessarily warrant an 88 three pointer for everyone.

It's mathematics. Contested shots have to do with Shot In Traffic which is still a probability driven rating. Shoot Off Dribble had a place in 2K games as well which was also probability driven.

There are also many objective and data-driven ways to determine things such as assisted 3PT FG's vs. non-assisted 3PT FG's, contested 3PT FG%, and using 3PT FG shot chart data to determine where the player is shooting from since most assisted 3PT FG come from the corners where the FG% is the highest.

Further more, one season of production shouldn't be the standard for a veteran player who has an entire resume of data. That's a shallow analysis.

The common mistakes that an unaware forum member makes when bantering about a rating should not apply to the professionals with access to the exact logic and probabilities behind the game. That is the point that I was making. None of the people discussing individual attributes in these threads have access to that info in 2K13 since it's not even completed yet.

[iLLosophy]

Quote: Originally Posted by youALREADYknow

The common mistakes that an unaware forum member makes when bantering about a rating should not apply to the professionals with access to the exact logic and probabilities behind the game. That is the point that I was making.

I understand what you are trying to say. The point I was making was, based on some of the ratings we saw in 2k12, and myself playing a year watching the types of jumpers people made in 2k12 last year, I don't think the person making the ratings fully understood the ramifications of giving Marreese Speights an 89 mid range jumper, at least when a human is controlling him. With the CPU, it's probably fine because of his tendencies will prevent him from going HAM since he's not programmed to take Carmelo like jumpers, but with a User that wants to let his sack hang, he can get away with making fair amount of contested jumpers with Speights in 2k12. In real life if a defender is not within like 1-1.5 arm lengths of Speights the moment he spots up (usually when he catches the ball), he's VERY money. But if a guy is within arms reach when he squares up, his shooting % falls dramatically (as with 95% of the rest of the respectable shooters in the nba). This dramatic fall in fg % doesn't convert in 2k12 for a guy with an 89 mid range, even considering shot in traffic.


I am hopeful since the 2k guys spent a lot more time working on tweaks this year that they figured this out and found some type of balance. Also with the new sliders Chris Smoove talked about, there might be a way to fix the issue i'm talking about on our own if it's still a problem out of the box.