Ratings/Tendencies for every player ever to play in the NBA, every year they played!

[Crossover1]

Title says it all. Not my work at all. As requested, big shout out to Fifedawg from NLSC for making this spreadsheet. This kind of work is very appreciated and will come in very handy for creating old teams for historic rosters. This is from his thread over on NLSC...

Quote:

The Spreadsheet is completed and ready for release. Every attribute in the game as been rating except for DUNK, STANDING DUNK & VERTICAL. These attributes cannot be rated for each individual players based on numbers so I'll leave those up to you guys. There are more than a few Tendencies that are not rated as well. Most of those I've left up to the user as well. Those include PUTBACK DUNKS, all under SET UP DRIBBLE, DRIVE RIGHT, all under DRIBBLE PENETRATION, all under FINISHING MOVE, all under SHOT TYPE, all under INSIDE SHOT, Some under VARIOUS SHOTS, only GIVE AND GO under FREELANCE, all under POST MOVES, only THROW ALLEY OOP under PASSING and only HARD FOUL under DEFENSE. I have my own style in doing these tendencies so if you want that info just let me know and i'll pass that on to you. I've also included the players real stats for there season in the spreadsheet as well. Some statistics may be marked as "N/A" and that simply means that there was no number for that particular statistic for that year. I've worked really hard on this for the past two months or so and appreciate your feedback. If you do use the ratings, all i ask is a simple "Shout Out." Happy Gaming! http://www.sendspace.com/file/e5ter5 Version 1.1 with 3pt Rating Fix http://www.sendspace.com/file/4acf0c Version 1.2 with Shot Ratings Fix http://www.sendspace.com/file/3lg33t My way of doing remaining ratings & tendencies http://www.sendspace.com/file/w8z4ls Version 1.3 with Updated 2010-11 Stats as of 1/20/11 http://www.sendspace.com/file/ux9n7h

[Sellaz32]

This has been discussed before and while I am quite impressed with the amount of detail and work that went into this project, I wonder how much these ratings translate to realistic gameplay. It doesn't matter how accurate the computations are, if they don't play out in the game.

Secondly, I believe there is a certain "art" to rating players that can't be applied to formulas. For example, Hal Greer was considered to be one of the best mid-range shooters in NBA history. He shot 45% from the field most his career (which was pretty average) Now when I create him, I am gonna give him a relatively high mid-range shot rating. Going strictly by numbers he gets lost in the pack, but if I am playing against him, I want to have to focus on stopping that mid-range jumpshot. (Not saying go crazy and give a 99 or anything)

Lastly, and I mentioned this before, but Magic's flashy passing rating was 24....Larry Johnson's was 27. I am having a hard time getting that one.

Ok, I am done criticizing now. I was just initially very excited about it, and yet disappointed when I actually saw some of the numbers.

Others may have different opinions about it.

Sellaz

[ashamael]

Quote: Originally Posted by Sellaz32

For example, Hal Greer was considered to be one of the best mid-range shooters in NBA history. He shot 45% from the field most his career (which was pretty average)

See, this is the problem with formulaic ratings over the entire history of basketball... You have to normalize stats due to massive changes in the game. Field goal percentage is a classic example of this. Look at the league average for fg% over the 60s. Then look at it over the 80s. Then look at it now. Some stat engines have normalized to a thing called fg+. How far above the league average was a player for the season?

Let's go with the great 66-67 Sixers team, for example. That year, Greer shot 45.9% from the field. The league average was 44.1%. That's 104% better. So if you take that same idea in the 85-86 season (which has a league average of .487), you're talking about someone shooting 50.6% from the field. Use that same idea last season (.461) and you're talking a 47.9% shooter. So which of those do you base rating him on? What he really shot in a different game? Or do you use the comparison and normalize it? If 2k had today's players rated accurately, I'd say it'd be simplest to just normalize everyone to today's ratings... but they didn't really do a good job with that, now did they?

Same thing happens with rebounds. Wilt & Russell's rebounding numbers were unreal. Oscar and his triple double? Absolutely ridiculous. But the pace of the game back then was much, much faster. More possessions (and more missed shots) gave more opportunities for players to grab boards. This makes it much harder to determine how to balance it out. If the NBA had counted offensive boards, we'd have some way of normalizing it with oreb% & dreb%, which would be the best way of doing so (Moses & Rodman, for example, are known as two of the best rebounders in the history of basketball... and when you look at their advanced rebounding stats, those numbers back up the claim)... but offensive boards weren't counted back in Wilt, Russell & Osscar's day, so anything you try is guess work... You could maybe try some kind of pace adjustment, but again, what is the standard to which you normalize?

[DaJuan]

Shots ratings are totally way off, it's a pity.

[Crossover1]

Yea I get what u guys are sayin. I was excited when I found this but when I was goin thru the ratings, many of them were very questionable. Kobe a better ball handler than AI? Not in a lifetime. And big guys had higher layup ratings than most guards. This whole thing is appreciated but I could see a lot of people being turned off by some of the ratings.

[Dice]

When I saw 87 Kareem's Off the dribble rating at 90, I was immediately turned off by it.

[CX1329]

I was turned off by it when I saw that 95-96 Shaq was given a 50 rating for 3PT shots because he made...1 out of 2 3 point attempts that season. Would he have been given a 99 if he had made both shots?

Sometimes, formulas can lead to nonsensical results like that. No disrespect, it must have taken copious amounts of work, but I don't think the formula worked.

[Sellaz32]

Which begs the question...wouldn't you take time to test this stuff BEFORE entering in all that data?

Wouldn't it also be cool if some of the "expert" tendency/rating guys had the time to share their principles more explicitly so someone like the guy who had the time to enter in all that data could apply a philosophy that works to every player? Essentially teaching how to fish rather than giving fish. Of course, there is less ego-gratification that way...but it would cut down on new guys like me spending hours testing and trying to dissect "what works." Or spending way too much time searching the forums for hidden clues on how to improve the gameplay. Maybe there isn't a set formula for what works....or maybe the 2K guys should be a little more clear about the effect different tendencies and ratings have on gameplay....

Sorry for the tangential rant, its just that I would rather be able to apply someone else's successful work to my own creation than waste time reinventing the wheel OR waiting anxiously for them to release their latest update.

No disrespect intended to any rating gurus out there. Their work is much appreciated!

Sellaz