The Warriors are heading into one of the most critical offseasons in recent franchise history, and frankly, it’s going to be messy. Right now, only four players are locked in for next season: Steph Curry, Brandin Podziemski, Gui Santos, and Will Richard. Everyone else is either dealing with injuries, has a player option to exercise, or is about to hit the free agent market. It’s chaos, and the front office has some serious decisions to make.

The two biggest questions revolve around center Kristaps Porzingis and guard De’Anthony Melton. Porzingis has been relatively healthy lately, he’s averaging 18.1 points per game in March and even called the Warriors’ director of sports medicine “the GOAT”. But his injury history and POTS diagnosis mean teams probably won’t offer him anything longer than two years. Melton, on the other hand, has made a solid comeback from his torn ACL and will almost certainly decline his $3.4 million player option to test free agency.

The problem? The Warriors might not have enough cap space to re-sign both guys. So here’s what we did: we built an AI tool called KRISTAPS (Knowledge-Rich Intelligent System for Tracking and Analyzing Player Salaries) to predict what these contracts might actually look like.

Yeah, you read that right. A journalist with zero coding experience, a laptop, and a whole lot of caffeine at Caffe Trieste in North Beach created a machine learning model to forecast NBA contract values. The tool factors in player statistics, injury history, age, and league trends to spit out predictions.

So what did KRISTAPS think? Porzingis could land a three-year, $37.5 million deal, roughly $12.5 million per year, which would eat up most of the Warriors’ non-taxpayer midlevel exception. Melton’s projection came in lower at three years and $28.5 million, or about $9.5 million annually, which tracks with current market rates for complementary guards.

Podziemski’s situation is interesting too. The Warriors could offer him a rookie extension this summer to avoid restricted free agency headaches down the line. KRISTAPS predicted a four-year, $74 million pact with $18.5 million annual average value, reasoning that his elite three-point shooting, defensive versatility, and proven health make him worth that investment.

Now, here’s the thing: KRISTAPS isn’t perfect. It projected LeBron James at just $2.5 million (lol), and it was way too conservative on Amen Thompson’s potential rookie extension. But for the Warriors trio, the predictions seemed surprisingly reasonable.

Obviously, actual NBA teams use way more sophisticated models than our passion project. And Mike Dunleavy and the Warriors’ front office will make real-world decisions that go way beyond what any algorithm can tell us. But sometimes, data can help us understand what’s actually possible in this weird, expensive world of NBA contracts.