Five Stats For Five Free Agent Signings
Plus a refreshing Q&A with the creator of Basketball GM
With the first day of free agency in the books, I thought it’d be fun to rip through a handful of free agent signings that caught my eye. I’ll focus on players that changed teams and provide some stats1 that help contextualize why these players got paid.
Luke Kornet - 4yr/41 million to San Antonio
Loyal readers know that I am Tier One Luke Kornet Truther, so it should come as no surprise that I am leading off with a chart highlighting Kornet’s superpower, which is his rim protection ability.
Last season, Kornet finished seventh in opponent rim field goal percentage when he was the closest defender. His new teammate, Victor Wembanyama, finished third. Lord help any opposing team facing off against a double-big alignment featuring Kornet and Wemby.
Dorian Finney-Smith - 4yr/53 million to Houston
Despite averaging less than ten points a game this past season, Dorian Finney-Smith posted a 1-Year Regular Adjusted Plus-Minus of +6.1 — meaning his teams were 6.1 points per 100 possessions better when he was on the court vs. off after controlling for the quality of his teammates and opponents. The only players that had a higher 1-Year RAPM value in 2024-25 were Shai Gilgeous-Alexander (+8.3), Nikola Jokic (+6.9), and Giannis Antetokounmpo (+6.7).
RAPM is good for identifying which players help you win games, but it’s not so good at telling you why they help you win. So here’s one reason why: DFS was one of the league’s best pick and roll partners this past season.
Among players that set at least 300 direct screens in the regular season, DFS finished 3rd in points per pick set (Kevin Durant, his future teammate, finished first). It helps that DFS’ primary pick and roll partners were Luka Doncic and LeBron James this past season, but any screener that can knock down 40 percent of their threes, like DFS did this past year, will make life easier for pick and roll ballhandlers looking for a reliable release valve. Amen Thompson and Fred VanVleet never had it so good.
Nickeil Alexander-Walker - 4yr/62 million to Atlanta
What if your best defensive stopper also had a reliable outside shot? That’s the pitch for Nickeil Alexander-Walker, who has quickly become one of the league’s most outstanding 3&D wings. This past year, opposing teams scored just 0.90 points per pick when NAW was the primary ballhandler defender. That ranked 3rd among players that defended at least 1,000 pick and rolls in the regular season.
Unlike a number of players on the above list, NAW also has a three-point shot that defenses will respect. He made 42.5 percent of his catch and shoot threes in Minnesota, which should translate nicely in Atlanta where Trae Young led all players in three-point assists this past season.
Ty Jerome - 3yr/28 million to Memphis
The floater has become the weapon of choice for ballhandlers that lack the juice to get all the way to the rim. No one knows this better than Ty Jerome who is yet to register a dunk in the NBA despite being 6'5". Last year, Jerome shot 52.3 percent on self-created floaters, which ranked third in the NBA behind Naji Marshall and Tyler Herro. It’s fitting he ends up in Memphis, which has been Team Floater going back several years now.
The Indiana Pacers exposed Jerome in the playoffs for the defensive weak link that he is, but he is still a valuable regular season player when he’s making shots. Also, he’ll get to run the show when Ja Morant is off the court, which, if history is an indication, will be often.
Jake LaRavia - 2yr/12 million to Los Angeles
The Lakers signing Jake LaRavia should take some of the sting out of losing Dorian Finney-Smith for nothing. This past season LaRavia cashed 45.7 percent of his threes from Above The Break, the best mark in the league among players that attempted at least 90 such shots.
His biggest problem is that he didn’t take enough. LaRavia attempted just five threes per 100 possessions this past season, which is roughly the same amount Evan Mobley attempted. If LaRavia wants to make an impact beyond what is expected of a $6 million a year player then the first order of business is to double his three-point volume.
A Refreshing Q&A with Jeremy Scheff, creator of Basketball GM
I get bummed out every year around this time when there are no more games to watch and most of the free agency has already concluded. Suddenly I have more free time than I know what to do with. I tell myself that I should use my newfound free time to do something unrelated to basketball. But truthfully, I just want to play Basketball GM.
Basketball GM is a simulation-based basketball management game that you can play in your browser. Jeremy Scheff released the first version of the game back in 2008 and has been incrementally improving it over the years. It doesn’t have all the bells and whistles of something like Football Manager, but it’ll more than scratch the itch if playing basketball on a spreadsheet is your idea of a good itme. Plus it’s free, doesn’t have any microtransactions, and it comes with an active online community.
I talked to Scheff about what makes Basketball GM unique, the role of AI in video games, and what he’d do if he were put in charge of the Sixers.
This Q&A has been lightly edited for length and clarity
F5: What makes Basketball GM unique from other GM simulation games?
JS: First of all, it's a free game (no microtransactions!) that runs in a web browser so you can play it anywhere. I think that already makes it unique.
In terms of gameplay, I'm going for something on the simpler side, with as little micromanagement as possible. I never enjoyed games that forced you to do stuff like manually click some button to scout a draft prospect or tell your player to work on his jump shot.
Despite the relatively simple core gameplay, there are a ton of features and options. Way more than you'd expect in a free web game. So overall I think the game is in an interesting niche... easy to pick up, minimal micromanagement, but an absurd amount of functionality built on top of the simple core.
How do you strike the balance between making a game feel true to life while also making sure it still feels like a game?
Mostly by leaving stuff out if it doesn't sound fun. For instance, contract rules... I know some people would like a super complex game with all the crazy contract and salary cap rules you see in real life. But I think most people (although maybe not most F5 readers!) prefer not to know every little detail of the CBA.
So I have min/max salaries, rookie contracts, Bird rights, a salary cap, and a luxury tax. No other exceptions, no player/team options, no incentives, no second apron, etc. I think that's about the bare minimum to feel somewhat realistic without having too much complexity.
That also helps when making a game because designing an AI that can intelligently navigate all those real life rules would be a lot more difficult.
How well do the rules in Basketball GM mirror the rules of the actual NBA? In other words, are there things that can happen in Basketball GM that cannot happen in real life (and vice versa).
By default it's pretty similar to the NBA, with some simplifications like described above. But there's also a lot of settings to change things. There are options for all the major rule changes that have happened in the history of the league (such as the 3 point line), plus some stuff that should happen in the future (Elam ending!), and some stuff that probably shouldn't (deciding tied games with a 3 point contest, similar to shootouts in hockey/soccer).
There's also a few areas where I have consciously not implemented things the NBA does that I don't like. The NBA eventually followed my lead and got rid of that "division winner gets a top 4 seed" rule. I tried to ignore the play-in tournament, but eventually I caved on that. I am still holding strong on the in-season tournament though!
Something I’ve thought about is how the impact of a rule change or a CBA change aren’t obvious until years down the line. I feel like there’s a way for a really true to life Basketball GM game to help us see the impact of those changes more quickly. Am I way off base?
Maybe.
Somebody has to code the AI for the game, and that will have a lot of assumptions in it. Maybe a multiplayer game, with human GMs controlling every team, could solve that problem. But then you still need a player AI making decisions. Like how sometimes a player's current team can offer a bigger/earlier contract extension than other teams in free agency, how much will players value that? A game can't really tell you that, it'd need some analysis done outside the game so the logic could be programmed in.
Then again, simulations do get used in many industries to help plan things, so probably there is some potential here. It might not look like a video game though.
You’re always tinkering with the game -- adding new features or making it more realistic. Do you have anything in store that you can talk about?
The big thing I want to do is to fix a lot of ratings/stats issues. If you play Basketball GM long enough, you will notice some things about player generation, development, and stats that seem a little bit off. I want to do one big fix to solve all of those problems, because they are all interconnected so it's difficult to make incremental progress.
But big changes are hard, so I've been procrastinating on that. Maybe talking about it here will motivate/shame me into making more progress.
How, if at all, has AI enhanced a game like Basketball GM?
AI in video games is a very controversial topic, and a lot of people are reflexively opposed to it. But I think sports simulation games are unique in how many opportunities there are for AI to enhance gameplay without compromising anything. Like...
AI-generated profile pictures for fake players. Heck you could even make them videos, not pictures!
None of that messes with gameplay, it's purely additive.
For a variety of reasons I have not actually integrated any of that into Basketball GM, but I would like to some day. In the meantime, some users are doing it on their own, like in the links above.
If you were given unlimited resources to improve Basketball GM, what would you work on?
Some big things that would be cool:
All the AI features mentioned above
Real multiplayer support
Multiple leagues within one save (like minor leagues or competing pro leagues)
A college version
A way to visually watch games live, rather than just seeing a play-by-play
A dedicated data science team constantly working on improving the realism of the simulation and rosters
At least some of those things will eventually happen even without unlimited resources!
Let’s say you were named the GM of your favorite team, what’s the first thing you’d do?
I'm a Sixers fan, so that is a very difficult question!
Embiid+PG+Maxey is basically the whole salary cap, so there's not a lot of flexibility unless you trade one of them (although Elon Musk might claim that DOGE can get them under the salary cap by cutting waste in 2-way contracts).
I think my first move would be to do nothing. Desperately trying to find some super clever trade to fix things might cause more harm than good. There's probably a better chance that Embiid and Paul George get healthy.
What’s one thing you cannot live without during the NBA season?
Can't decide between Basketball Reference and DARKO.
For more interviews, visit The F5’s Q&A archive. I’ve done interviews with the creators of EPM, DARKO, LEBRON, xRAPM as well as ESPN’s Dean Oliver, Kevin Pelton, and more.
All stats via Genius Sports
"His new teammate, Victor Wembanyama, finished third."
Yes but his old teammate finished 4th...
Luke Kornet, undercover menace. I really do think the Spurs are going to be so fun/so scary so so soon.