Star ratings vs actual player value

[emperorsnapper]

One thing I have noticed in prior versions of this game, that continues in 2021, is the way that the star ratings are heavily influenced by certain player attributes, while ignoring others to the point where they become a poor representation of player quality.

The essential issue is this: star ratings are heavily influenced by a player's "scoring" rating, while the "defense" rating has little effect.

The thing is, a high scoring rating is sometimes not even a positive attribute. There is no question that the worst players in DDSPB are those with a high scoring rating, but low FG%. These guys will just kill your team, going 10/30 every night. Pair that with a poor defense rating and you have a guy who might have 4 or 5 stars, cost $30 million a year, and doom your team.

On the other hand, a player with a high defense rating and a low scoring rating is always helpful, even when they are not good shooters (because they won't kill you by shooting too much).

Complicating this a bit is the fact that the FGI and FGJ ratings do not necessarily translate into decent FG%. I have seen guys with a FGJ of 5 consistently shoot well from deep, while others with an 8 or 9 shoot 25%. The best way to tell who can hit shots is to look at their stats over their ratings (which is a bit more difficult for young players).

As an experiment, I took the Sacramento Kings and edited the entire roster. I kept all ratings and traits the same, but for 2: I turned everyone's scoring rating down to 20, and everyone's defense rating up to 100. Shooting, stealing, etc. all remained whatever they were originally.

Most of the team wound up with 1 star overall ratings; some had 1.5. DeAaron Fox was the only player with 2 stars, likely because of his great passing/handling skills. According to the star ratings, this was the worst team in NBA history. Utterly doomed, and not likely to compete in even the G-league.

In 51 games with this modified roster, they went 26-25. They were the lowest scoring team in the league, but also allowed the fewest points. They were not a good shooting team(but the Kings are not a great shooting team anyway), but the biggest problem with the lack of scorers was the team would take forever to get a shot up. There was no pushing the pace because no one wanted to score, but they were still decent, about as good as an unedited Kings team would have been.

I then re-edited the roster. I turned everyone's scoring up to 100, and their defense down to 20. The lowest star rating was now 3 stars. Fox and Holmes were 5 star players. No team came close to such star quality - they appeared to be the class of the league. They went 9-22 the rest of the season and missed the playoffs.

Scoring went up, points allowed exploded. The team FG% didn't improve at all, but they took a lot more shots. Meanwhile, they could no longer stop anyone and just got whipped most nights.

Obviously, there are issues with this experiment - there are diminishing returns to having great scorers - there is only one ball, and a 5th option who has a scoring rating of 100 isn't really doing anything for you. OTOH, every strong defender provides roughly equal additional value - having 5 on the floor at all times means there is never an easy matchup to exploit. Also, the Kings are not a great shooting team, which hurts them in every case, but hurts more when they are maximizing the number of possessions by shooting a lot.

Still, the fact is a guy with a great defense rating who hits his shots is a valuable player even if his scoring rating is low, but such players almost never get more than 2 stars and are often available for league minimum. Meanwhile, a player who shoots poorly but has a high scoring rating is a team killer (especially with a low defensive rating), and they demand high salaries to lose games for you. Focusing on collecting strong defenders with solid FG% creates a huge opportunity for arbitrage, and once you now this it is unfortunately far too easy to build a team that consistently wins ~70 games a year.

[Gary Gorski]

Let me address two different things in your post here. First is the "star rating" system - the star rating system is NOT an overall look at how valuable a player is to a team. It monitors reality - let's face it, if you don't put the ball in the basket nobody considers that player to be a star. Marcus Smart is probably one of the best defensive players in the league but for his career he averages like 10 PPG so he's not a "star". Carmelo Anthony was considered an all-star for most of his career yet his teams never won anything in the NBA but he was still a "star"

The good news, as you can see, is that the results of the game are based on the individual ratings and not the "star system" which means you actually have to find players that best fit a system and not just try to collect the highest "star" players. Now your experiments weren't exactly realistic but your assignment of "worst team in NBA history" was your perception of that based on the star rating system which as you can see the game did not produce the worst team in NBA history and instead produced what you might expect of a team with a roster full of the greatest defensive players of all time who had extremely limited skills on offense. By the same token when everyone was an offensive hero but played no defense as you suggested the returns of doing that offensively were far diminished and the lack of defense hurt far more and produced a worse team.

Your final conclusion that " a guy with a great defense rating who hits his shots is a valuable player even if his scoring rating is low" is correct BUT you left out the asterisk there where he is only valuable IF his role is to be the defender who hits open shots because he can operate as the 4th/5th option. If that guy is expected to be your number one offensive option then no, he's really not going to be that valuable because his skill set is only valuable as part of the piece of the puzzle, not as the singular ingredient to winning and yes a high scoring, low % guy is often times a killer - how many times has a player like that led a team to an NBA championship. I mean you're describing guys like Westbrook and Iverson on offense...yet they are considered "star" players and were amongst the highest paid in the league.

I'm not sure about your final comment that collecting strong defenders with solid FG% makes it easy to build a 70 win team. Your experiment with all 100 defensive players only turned out to be a .500 team. Now if you have a file with a team that is full of 2 star players that is a 70 win team I would love to look at it but I would be surprised to see that happen. Now would pairing some of those guys with one or two really solid offensive players who are efficient make for a great team? Yeah I mean if you have two star offensive players who are not chuckers and everyone else is playing excellent defense that's going to be a tough team to beat but again, that's realistic.

[brewers8_28]

I never in a million years would’ve guessed the stars represented actual star power.. like I think everyone assumes it’s based on overall talent. And the game doesn’t really ever explain what the stars actually represent. like the way it’s done in the college football game better, with numbered overall ratings. Who cares about measuring a players star power? How does that help me evaluate players? Why is it important for me to know that info? Wouldn’t an overall rating based on talent make more sense than an overall rating based on star power?

Im just whining though. Honestly your games are amazing, and it says a lot when the biggest flaws are little things like this.