Talkathon has a projected NBA 3pt% that is based off of some calculation related to 3pt%, FT%, and 3PA. I decided to see how predictive their tool is, and so I compared the projected NBA 3pt% with the actual NBA 3pt% of players from the last 5 drafts. I compared their first year 3pt%, 2nd year 3pt% and career 3pt% vs the projections. In total I looked at about 50 players.
- The average difference for both the first year 3pt% and second year 3pt% was about 3.5% off. So if a player was projected to have a 3pt% of 35%, their actual 3pt% would be on average between 31.5% to 38.5%.
- The average difference for career 3pt% was only 2% off, or in other words as more NBA 3pt data was available, players shot closer to their projections.
Anecdotally, I would say their tool is actually a very good predictor of 3pt%, but that the variance for most players is due to shooting more catch and shoot vs pull up 3pt shots in the NBA. It seemed like the prediction would be higher than actual NBA 3pt% for players that take a lot of pull up 3's in the NBA, and the prediction is lower than actual NBA 3pt% for players that take mostly catch and shoot 3's.