As discussed some last week, PGs have generally seen a boost to their assist numbers and a slight drop in their turnovers compared to their career figures when playing for Flip. I'm not sure you can say that Flip "changes" his PGs in other ways. Which might be for the best. Some more discussion of Flip's past PGs, a comparison to the EJ PG experience, and some predictions for Flip's impact are below.
One might guess from Flip's reputation that his PGs would post unusually high numbers of 3s. No, not really. That's probably mostly just a product of Billups' skill set, which has always involved a lot of 3-pt attempts. Cassell, never shy to shoot, shot half as many 3s as Billups-those figures are right in line with their respective career numbers.
One might guess from Flip's reputation that his PGs would post unusually low numbers of FG attempts, spending more time setting up their teammates. No, not really. Every Flip PG but Terry Porter has had at least one season with Flip above their career average in shot attempts, and Sam Cassell, the most "shoot first" of his PGs, actually took more shots in his time with Flip than his career average. (Given the personnel on those teams, it isn't really surprising, but still.) Billups had seasons with above-average FG attempts (and below average) during both of his stints with Flip.
Flip has had very different players to work with at the point in his career. They've been diverse ages (the end of Terrell Brandon's career vs the beginning of Starbury's) and sizes (6'1" 180 lb Troy Hudson versus 6'3" 202 lb Chauncey Billups), and have had a wide range of usage rates, shooting efficiencies, etc., both in and out of Flip's offense. There is no such thing as a "typical" Flip PG, and it is not clear that he has made any effort to get guys to significantly change their style.
Bottom line: he'll maximize a PG's "pg type" stats (as doclinkin had noted), but that by no means necessarily comes at any cost to the shooting stats. It seems, if anything, Flip has sought to take advantage of his PG's shooting skills to fill in around what the team otherwise needs. For example, Cassell took extra shots on a team that otherwise lacked in shot creation (while not increasing 3-pt attempts, as Sprewell and Hoiberg were doing enough of that), and Billups took lots of 3s on a team without a classic 3-pt shooting SG/SF.
Moving away from the EJ-Princeton
Of course by now it has been discussed a lot that the switch from the EJ-Princeton to a Flip offense will change the role of the PG in the Wizards' offense. It is difficult to analyze statistically exactly what the EJ-Princeton does to a PG's production because he has had so few as a head coach.* We don't have much pre-EJ data on Arenas, of course, and while Antonio Daniels also accumulated a significant amount of run playing for EJ, I'm not sure we can put much weight on that data.**
For what it is worth, however, it is interesting to note that Arenas had the highest assist rate of his career the year before coming to Washington. In his age 21 season in Golden State, he posted a rate of 30%, with 6.5 assists/36 minutes, which has also been a career high. It really isn't normal for a PG to peak as a distributor that early in his career. Two players who had seasons similar to Gilbert's at such a young age, Baron Davis and Tony Parker, both went on to have 40%+ assist rates in some later seasons. Even Iverson later posted some seasons solidly in the 30s. Gilbert, in contrast, has not been back above 27.2%, which is much more Steve Francis-like as PG career progressions go.***
Gilbert turned the ball over more in Golden State than he eventually did under EJ, but the signs point to normal NBA-PG maturation as the cause, since it took a couple of years in Washington to get the TOs under control.
In any event, the data is consistent with the perception that EJ did not ask his PGs to be PGs in the conventional sense. (Note that this is not itself intended as any criticism of EJ per se, just an observation of the impact his system has at the PG position.)
Moving, therefore, straight from a EJ-Princeton to a Flip offense may have nearly as much of an impact on a PGs' assist numbers as one will see. (I suppose going to the D'Antoni system would make for an even bigger change.) And, it won't necessarily hurt the player's scoring, etc.
Predictions
As the excitement surrounding Arenas's return continues to build, this certainly doesn't do anything to dampen it. Thanks to the likely drop-off in minutes, if you focus on pure per-game counting numbers, this might not end up looking like Arenas's best season (unless the team plays more up-tempo than expected). But, by most other measures, Flip and Arenas make for an exciting combination. I'm predicting that a healthy Arenas posts the best "pg" stats of his career, with a potentially huge increase in assists without any deterioration in turnovers, while maintaining his shooting efficiency and largely maintaining his shooting and usage. (Although, given the personnel changes, we could see a drop in the proportion of his 3-pt attempts.) If Arenas is healthy this season, I'll be surprised if he doesn't post career highs in PER and Win Shares, even if he isn't quite the pre-injury Agent Zero. If he is....look out.
While I recently argued in comments that we shouldn't assume Deshawn can't compete for backup PG minutes, I have to say I'm very interested to see what Foye might do as a PG on this team. He has already displayed some early-Sam Cassell-like abilities. (Differences being that Foye has posted fewer assists but is a better outside shooter.) If Flip decides to use him as the backup PG and can perk up Foye's assist/TO numbers, Foye could prove to be a very valuable backup PG. Perhaps one of the best in the league? Foye has already cut his TOs to acceptable levels. His assist rate has been a flat 20%, however. But now he will have better weapons to pass to than ever before. By a long shot. This is could be a huge opportunity for him. Whichever of he and Miller end up as the "6th man" on this team should end up in the discussion for 6th man of the year.
Crittenton is harder to get excited about from a stat perspective. The best that can be said is that he has a lot of room for improvement, and it must be noted that he is still very young. But, among Flip's past PGs, at this point he compares unfavorably to a young Troy Hudson. Foye has posted much better distributor numbers than Crittenton has (though he does have a significant age advantage.) Crittenton might need to make a lot of progress in a hurry to get opportunities here.
* There isn't much to be done with EJ's time as an assistant in NJ either, I think, since Kidd is probably a special case. Kidd did post 40+% rates while there.
** Antonio Daniels saw his assist rate dip a bit when he got to Washington, but he was also inconsistent throughout his career (playing different roles in different career stops) and was entering his 30s, so it is hard to draw conclusions.
*** This all leads me to wonder whether a lot of Arenas's extreme "shoot-first" reputation came about because of being the PG in an offense that doesn't really call for a PG. Can't help but wonder how different his career line (and reputation) would be by now if he'd played in a "normal" offense through his mid-20s. Of course, Davis and Parker are both considered shooting PGs as well, but my impression is the NBA media takes them more seriously as distributors than it does Arenas. I'm sure Arenas still would have had a fairly high number of shot attempts, but the ratio of shots to assists might be lower. Maybe this season we'll find out. Billups never topped 30% assist rate until his second time with Flip, when he shot all the way up to 38%.