I’ve been reading Paul Knepper’s great new book on the 1990s Knicks. Paul was kind enough to give us a an extensive interview a few months ago and we examined the compelling story lines. Reading his book got me to wondering about looking back at some of the details of that run. How do the advanced stats of those teams compare with the prevailing narrative of the 1990s Knicks? I thought we could take a look. (By the way, you should definitely buy the book!)
The Brief Recap of the 90s Knicks
The broad story is pretty well-known but let’s run through the basics for those who happen to be a little younger and don’t remember a time when the Knicks were considered a model organization (yes, it’s true). The Knicks hired Dave Checketts as president in 1991 and he brought in Pat Riley to coach a frustrated Patrick Ewing. Riley turned the Knicks from a .500 team to instant contenders by getting the Knicks to play incredibly tough (and perhaps a bit too rough) defense. They rode Ewing and the D to nearly beat Michael Jordan a few times before losing in the 1993-94 Finals to Hakeem Olajuwon in a tough Game 7.
Thereafter, the core aged and Riley bolted town for Miami in 1995. It’s hard to conceive of now but the Knicks were a very attractive free agent destination and they had a shrewd front office, so they rebuilt on the fly. They had their choice of the free agent crop and got Allan Houston (free agent signing in 1996), Larry Johnson (traded for Anthony Mason in 1996), Marcus Camby (traded for Charles Oakley in 1998), and Latrell Sprewell (traded for John Starks and some other stuff in 1999). The Knicks came close again with the new crew but came up short before James Dolan took control and systematically crushed the organization. Dolan has run MSG like an incompetent dictator ever since.
Okay, with all that out of the way, here are a few fun detours I wanted to take in the story…
How did Riley improve the Knicks so much in one season?
In 1990-91, the season before Riley got town, the Knicks were 39-43 and were roasted terribly by Michael Jordan and the Bulls in the first round of the playoffs. In addition to several memorable facial dunks, the Bulls outscored the Knicks by 20 ppg for the three-game sweep. Ewing was annoyed at his bad team and wanted to go to a contender. How do you fix this quickly?
The Knicks had Patrick Ewing at his absolute peak (26.6 ppg 11.2 rpg, 3.0 apg, 3.2 bpg) but New York was 16th in offense and 12th in defense. The offense was not great. Too many bad shots by role players like Charles Oakley and Gerald Wilkins. A tongue-in-cheek description from the 1992 Rotissiere League Basketball Book wrote about this ugly offense with this (probably) apocryphal story: “[u]pon kissing the tarmac on U.S. soil, American POWs told harrowing tales of how their Iraqi captors, in violation of the Geneva Convention, forced them to watch films of Wilkins leading the Knicks fast break.” Maybe that’s a little much but it was fair to say that the Knicks were a slow-paced and bad offensive team and an average defensive team.
Anyway let’s look at the advanced stats of the key players for the 1990-91 squad:
PG, Maurice Cheeks: 14.2 PER, 5.2 WS, .117 WS/48, 1.3 BPM, 1.7 VORP
SG, Gerald Wilkins: 13.3 PER, 3.0 WS, .066 WS/48, -1.0 BPM, 0.5 VORP
SF, Kiki Vandeweghe: 14.4 PER, 5.7 WS, .112 WS/48, -0.8 BPM, 0.7 VORP
PF, Charles Oakley: 13.6 PER, 6.2 WS, .109 WS/48, -1.0 BPM, 0.7 VORP
C, Patrick Ewing: 23.7 PER, 10.0 WS, .155 WS/48, 4.2 BPM, 4.9 VORP
Key Bench Players
PG, Mark Jackson: 18.2 PER, 4.2 WS, .126 WS/48, 2.5 BPM, 1.8 VORP
SG, Trent Tucker: 11.7 PER, 1.9 WS, .078 WS/48, -0.5 BPM, 0.4 VORP
SG, John Starks: 14.3 PER, 2.4 WS, .097 WS/48, 0.9 BPM, 0.9 VORP
SF, Brian Quinnett: 9.0 PER, 0.8 WS, .039 WS/48, -3.4 BPM, -0.3 VORP
SF/PF, Kenny Walker: 10.3 PER, 1.3 WS, .082 WS/48, -3.4 BPM, -0.3 VORP
PF/C, Jerrod Mustaf: 8.3 PER, 0.4 WS, .022 WS/48, -5.0 BPM, -0.6 VORP
Yeesh…where to start? The guard situation, with the exception of Wilkins, looked solid. The obvious plan based on these numbers would be to give Starks more playing time than Gerald, who could potentially be useful off the bench. The bigger problem was the forwards. Oak, who is remembered for his toughness and defense, was actually a negative DBPM player every year of his career through 1990-91, and he shot too much. At least, Oakley was clearly a legitimate NBA starter. Kiki could still score a little but couldn’t score enough to justify his defense at his age. Finally, the bench forwards/centers were basically unplayable. Quinnett as the first forward off the bench? Mustaf was a rookie and as not ready to play PF or C at that point (or, ultimately, ever in the NBA). And we didn’t even bring up the deep bench centers Eddie Lee Wilkins and Stuart Gray, who were even worse.
Checketts and Riley seemed to understand all these points well. They made three moves that summer (in addition to placating an unhappy Ewing). First, they traded Mustaf for Xavier McDaniel. Then they drafted Greg Anthony to replace Cheeks (who left via free agency and immediately fell into a backup role). Finally, they signed Anthony Mason off the scrap heap and he absorbed all those minutes that Walker and Mustaf had not made the most of. Gerald remained the starter but Starks nearly equaled his minutes. Kiki remained but was shuttled into the Quinnett role and Quinnett was sent to the deep bench (Walker was cut loose).
The result: the Knicks went 51-31 with a 3.67 SRS. The offense improved a bit (12th) but the defense jumped to second in the NBA. Here’s how the advanced stats looked for 1991-92 Knicks:
PG, Mark Jackson: 17.8 PER, 7.6 WS, .149 WS/48, 3.6 BPM, 3.5 VORP
SG, Gerald Wilkins: 12.0 PER, 4.0 WS, .082 WS/48, -1.0 BPM, 0.6 VORP
SF, Xavier McDaniel: 13.9 PER, 4.6 WS, .093 WS/48, -1.3 BPM, 0.4 VORP
PF, Charles Oakley: 11.3 PER, 5.7 WS, .119 WS/48, -0.4 BPM, 0.9 VORP
C, Patrick Ewing: 22.8 PER, 13.0 WS, .198 WS/48, 4.9 BPM, 5.5 VORP
Key Bench Players
F, Anthony Mason: 11.3 PER, 5.6 WS, .123 WS/48, -0.8 BPM, 0.7 VORP
SG, John Starks: 16.4 PER, 6.4 WS, .146 WS/48, 2.6 BPM, 2.4 VORP
PG, Greg Anthony: 10.6 PER, 2.2 WS, .070 WS/48, -1.9 BPM, 0.0 VORP
SF, Kiki Vandeweghe: 14.2 PER, 2.9 WS, .144 WS/48, 0.6 BPM, 0.6 VORP
Riley got big improvements out of Jackson and Starks and had no real holes in the bench, compared the gaping holes of 1990-91 (funny side note was that Kiki was a decent bench player in 1991-92, which is something I totally did not recall). That was enough to move the Knicks into a second round playoff team. Not quite title contenders yet but closer.
X-Man v. Charles Smith
A prevailing sentiment of Knicks fans was that letting McDaniel go in free agency after the 1991-92 and replacing him with Charles Smith (acquired with Doc Rivers and other stuff in a three-team trade where they gave up picks and Mark Jackson) was a bad move because Smitty was not as tough a player.
Smith was part of that legendary end of Game 5 against the Bulls where he was unable to score over Horace Grant and Michael Jordan. What’s forgotten is that X-Man was pretty ineffective for much of the 1991-92 season. He started hot and finished great in the playoffs but he was an older player with bad knees. Here’s how the two players did in 1992-93:
-Smith (age 27): 14.9 PER, 6.6 WS, .146 WS/48, 1.2 BPM, 1.8 VORP
-McDaniel (age 29): 16.3 PER, 4.4 WS, .096 WS/48, -0.1 BPM, 1.1 VORP
Neither were overwhelmingly great. The Celtics gave X-Man a three-year deal worth about $7.4 million (plus two option years) to replace the retiring Larry Bird. The interesting back story is that the Knicks had a large amount of cap space ($14 million) and wanted to use it to get another key player before signing McDaniel. While the Knicks were attempting to sign Harvey Grant from the Bullets, Boston pounced and snatched McDaniel away.
Banking on Harvey Grant? Not a very inspiring scenario. Harvey was not a great defender and merely a decent scorer on a bad team. In any event, the Celtics bought a rapidly depreciating asset (X-Man was much worse in the second and third years in Boston) and the Knicks went with Plan B, Smith.
Smith was a better option on paper but he, also, did not fit the real need, which was an athletic small forward. A healthy McDaniel would’ve been perfect but he was never really healthy again. The bottom line is that swapping out X-Man for Smith was a moderate gain for the regular season but it’s hard to say that it made any difference (no, I don’t think X-Man would’ve scored either had he been in the same position that Smith was at the end of Game 5). In the end, neither Grant, X-Man, nor Smith was the right fit. There really wasn’t a pure small forward available at that time. The only good small forward that might’ve been available was Danny Manning of the Clippers, who was better than his teammate Smith and didn’t love being a Clipper. But Manning was too good to really be available. Rather, New York had all that cap room and no great way to spend it for that position.
1990s Knicks: Tough D and Slooooow O
Yes, the Knicks defense was quite good. To get a little context as to how the Knicks developed as a team in the 1990s, here’s how their year-by-year team advanced stats looked:
1990-91: 39-43, -0.43 SRS, 16th Offense, 12th Defense, 22nd Pace (Jon MacLeod coach)
1991-92: 51-31, 3.67 SRS, 12th Offense, 2nd Defense, 25th Pace (Riley’s first year, took Bulls 7)
1992-93: 60-22, 5.87 SRS, 22nd Offense, 1st Defense, 23rd Pace (Lost to Bulls after having 2-0 lead)
1993-94: 57-25, 6.48 SRS, 16th Offense, 1st Defense, 24th Pace (Lost to Rockets in Finals)
1994-95: 55-27, 2.78 SRS, 16th Offense, 1st Defense, 23rd Pace (Riley’s last year, Ewing misses finger roll against Pacers)
1995-96: 47-35, 2.24 SRS, 21st Offense, 4th Defense, 22nd Pace (Don Nelson coached, Jeff Van Gundy came in for last 23 games)
1996-97: 57-25, 3.31 SRS, 25th Offense, 2nd Defense, 15th Pace (Allan Houston and Larry Johnson come to team, lost to Miami after team most of the team was suspended for Charlie Ward/PJ Brown fight)
1997-98: 43-39, 2.74 SRS, 20th Offense, 4th Defense, 25th Pace (Ewing hurt most of the year, upset Miami in the playoffs after LJ/Mourning fight)
1998-99: 27-23, 1.45 SRS, 26th Offense, 4th Defense, 24th Pace (Latrell Sprewell and Marcus Camby come to team, make magical run to Finals)
1999-00: 50-32, 1.30 SRS, 21st Offense, 6th Defense, 29th Pace (Lost in Conference Finals to Pacers, Ewing traded after season)
2000-01: 48-34, 1.98 SRS, 19th Offense, 3rd Defense, 29th Pace (Van Gundy’s last full season as coach, lost in first round to Raptors)
2001-02: 30-52, -4.15 SRS, 25th Offense, 18th Defense, 21st Pace (Van Gundy quits 19 games into season and they finished 20-43)
A few things stand out from these numbers:
-The Riley/Van Gundy Run was uniquely focused on defense to such an extent that they never had an above-average offense besides Riley’s first year. Even the later teams with glamour offense players like Sprewell and Houston were below-average on offense. This was clearly by design. The slow pace prevented easy fast breaks the other way.
-I had always considered the 1992-93 team the best of the Knicks teams (and Jeff Van Gundy agrees). They had a best record and peak Ewing but SRS actually liked the 1993-94 slightly better. Of course, minors SRS differences aren’t a reason to conclude 1992-93 wasn’t the better team but it was interesting to note.
-Man, was Riley a great defensive coach. Leading the NBA in defense three straight years is impressive. In viewing DBPM, almost every regular player was a defensive positive during that time. The only player during the 1992-95 run who played a bit and who did go slightly negative in DBPM was Hubert Davis in 1994-95 (-0.2).
The Knicks’ strategy was to eschew all double teams so that there would be few easy open jumpers and it worked really well. The fascinating question that one wonders is whether the poor slow-paced offense was a necessary consequence of the defense. It’s hard to say but if the Knicks could’ve been slightly above-average on offense, they could’ve won a title. We’ll talk about this more below when we get into the point guard situation.
-Riley must’ve felt the ground quaking in 1994-95. On paper, the Knicks were only two games worse than 1993-94 and their offense and defense rankings were the same but the SRS collapsed and their expected win-loss was 50-32 (which is how they played against the Pacers in the playoffs that year).
-Van Gundy was a real Riley protégé strategically. The Knicks remained a slow-paced, defense-first team with him in charge. The 1996-97 team was particularly interesting because that was Ewing’s last year as a true star and the first year they brought in Allan Houston and Larry Johnson (John Starks and Charles Oakley were still there and still pretty good). It felt like that 57-win team could have a last showdown against the MJ Bulls. It didn’t happen because of the Ewing suspension against the Heat. Nevertheless, there isn’t a ton of evidence that the Knicks were anywhere close to the Bulls at the time. The Knicks significantly outperformed their won-loss expectation (50-32) that season. I would’ve loved to see the series but it would’ve probably been an easy Bulls victory.
-The Houston-Sprewell team up was fun but Van Gundy’s Knicks also won mostly with defense. Without Ewing for most of 1997-98, the team was still strong in defense and Charlie Ward was the best player by nearly every advanced metric (in fact his 3.4 DBPM led the NBA). Both Ward and Chris Childs consistently scored as strong defenders at the point guard slot.
The only Knick who consistently had a negative DBPM under Van Gundy was Houston (and he was closer to flat than a big negative). All the forwards and centers, particularly Camby, were quite impactful. The story of the success in the Van Gundy teams was his ability to get great PG defense and great defense from Camby. After Van Gundy left in early 2001, nearly all the players went negative defensively (or even more negative in the case of Houston).
Knicks PGs of the 1990s: clunky but effective?
In 1993, the legendary rap group A Tribe Called Quest came out with its classic album Midnight Marauders. In the liner notes for the album,, Phife Dawg, a huge Knicks fan, wrote that the Knicks would never win until they got a real point guard. Phife saw the Knicks point guards like most fans saw them, as limited. Indeed, the PGs were not smooth at all and could not consistently breakdown the defense. Take the group and how they (rightfully seen by fans):
-Mark Jackson: slow guy who was a good passer
-Greg Anthony: good defense, bad shooter/passer
-Doc Rivers: older and just okay at every facet of the game (though he wasn’t bad at anything)
-Derek Harper: older and not a great passer but a great defender (love that hand check!)
-Charlie Ward/Chris Childs: undersized guy who tries hard on defense but not a great offensive player
Ward and Childs were considered very lame by Stephon Marbury, who was not shy to criticize them. After losing a tough game to the Knicks in April 1999, Marbury kindly observed: “[the Knicks will] never win no championship with them two guys [Chris Childs and Charlie Ward] You can put that on the back page if you want to. They might not even make the playoffs.” Well, the Knicks did make the playoffs (and the Finals that season). To think that no one thought to pull Marbury aside and show him the DBPM stats that showed that Ward and Childs were really good! In any event, here are the advanced stats of all the lead PGs for the Riley/Van Gundy Knicks:
-1991-92, Mark Jackson: 2.7 OBPM, 0.9 DBPM, 3.6 BPM, 3.5 VORP
Greg Anthony: -2.7 OBPM, 0.8 DBPM, -1.9 BPM, 0.0 VORP
-1992-93, Doc Rivers: -0.6 OBPM, 3.6 DBPM, 3.0 BPM, 2.4 VORP
Greg Anthony: -0.6 OBPM, 3.3 DBPM, 2.7 BPM, 2.0 VORP
-1993-94, Derek Harper: -0.5 OBPM, 2.6 DBPM, 2.1 BPM, 1.4 VORP (54 games)
Greg Anthony: -0.8 OBPM, 2.4 DBPM, 1.6 BPM, 1.8 VORP
Doc Rivers: -0.1 OBPM, 2.3 DBPM, 2.2 BPM, 0.5 VORP (19 games)
-1994-95, Derek Harper: -0.1 OBPM, 0.3 DBPM, 0.2 BPM, 1.5 VORP
Greg Anthony: 0.9 OBPM, 1.7 DBPM, 2.5 BPM, 1.1 VORP
-1995-96, Derek Harper: -0.2 OBPM, 0.7 DBPM, 0.5 BPM, 1.8 VORP
-1996-97, Chris Childs: -1.4 OBPM, 1.0 DBPM, -0.3 BPM, 0.9 VORP
Charlie Ward: -1.8 OBPM, 2.2 DBPM, 0.4 BPM, 1.1 VORP
-1997-98, Charlie Ward: 0.7 OBPM, 3.4 DBPM, 4.1 BPM, 3.6 VORP
Chris Childs: -1.7 OBPM, 1.5 DBPM, -0.2 BPM, 0.7 VORP
-1998-99, Charlie Ward: -0.8 OBPM, 2.6 DBPM, 1.7 BPM, 1.5 VORP
Chris Childs: -1.8 OBPM, 1.3 DBPM, -0.5 BPM, 0.5 VORP
-1999-00, Charlie Ward: 0.3 OBPM, 2.5 DBPM, 2.8 BPM, 2.4 VORP
Chris Childs: -2.9 OBPM, 0.5 DBPM, -2.4 BPM, -0.2 VORP
-2000-01, Charlie Ward: 0.6 OBPM, 1.5 DBPM, 2.1 BPM, 1.5 VORP
Chris Childs: -3.1 OBPM, 1.2 DBPM, -1.9 BPM, 0.0 VORP (51 games)
Mark Jackson: -1.2 OBPM, 0.5 DBPM, -0.6 BPM, 0.3 VORP (29 games)
The above review tells us the Knicks were usually adequate-to-good at point guard. The 1991-92 version of Jackson and 1992-93 Doc Rivers were very good and Harper was solid. Ward was surprisingly effective (even though the eye test didn’t show him to be super effective). Childs and Greg Anthony struggled on offense but were pretty good defensively. Sure, Marbury was a better player than all of them but not by nearly as much as the raw stats indicate, primarily because Marbury’s defense was mostly terrible (-1.4 DBPM).
Before we dismiss the notion that the Knicks could’ve been much better at the point slot, though, we have to look at Tim Hardaway. Like Marbury, T-Hard was a very talented offensive point guard who gave up a bunch on defense. When he nabbed Hardaway for the Heat in 1996, however, Riley was able to get Hardaway to play some defense while also maintaining the same offensive production.
So, it does seem that the Knicks probably made a mistake in settling for Harper in 1994. He was a fine player but if the Knicks had been a bit bolder and tried to get a player with more offensive upside, it would’ve served them well. To be fair, Hardaway didn’t really come available until after the core of the Riley Knicks had declined. The intriguing gamble for the Riley Knicks would’ve been Rod Strickland, who was available for peanuts at various times in the early 1990s because he was not easy to deal with (the Spurs let him walk for nothing in 1992). Would a healthy and committed Strickland been enough to push the Knicks past the Bulls/Rockets? We can’t say for sure but it would’ve been fun to see (but still probably not).
How good were the Knicks versus non-Bull teams?
If you pretend the Bulls didn’t exist in the 1990s, the Knicks are in the conversation regarding the best teams of that time. The 1993-94 Knicks were the best non-Bulls team in the Eastern Conference by SRS (though the 1994-95 Magic are virtually tied). The issue is that the west had several teams with better SRS than the Knicks. The Sonics, Blazers, Lakers, Jazz, and Spurs all had multiple teams rated better than the Knicks. The 1993-94 Knicks rank 21st in SRS among all teams over that time span (New York would be 16th if you exclude the Bulls), which suggests that New York needed a whole bunch of things to go right to win a title. Alas, it was the Rockets (who scored as worse in SRS than the Knicks) who got everything to break right and took the non-MJ titles of the 1990s.
All-90s Knicks Team
Finally, to sum up, here are the best Knicks seasons, by position, of the 1990s squads:
PG, Charlie Ward 1997-98: 28.3 MPG, 7.8 PPG, .455 FG%, 3.3 RPG, 5.7 APG, 15.6 PER, 7.1 WS, .148 WS/48, 4.1 BPM, 3.6 VORP
SG, John Starks 1992-93: 31.0 MPG, 17.5 PPG, .428 FG%, 2.6 RPG, 5.1 APG, 16.3 PER, 7.5 WS, .145 WS/48, 2.8 BPM, 3.0 VORP
SF, Anthony Mason 1994-95: 32.4 MPG, 9.9 PPG, .566 FG%, 8.4 RPG, 3.1 APG, 14.9 PER, 8.6 WS, .166 WS/48, 1.9 BPM, 2.5 VORP
PF, Charles Oakley 1996-97: 35.9 MPG, 10.8 PPG, .488 FG%, 9.8 RPG, 1.4 APG, 14.2 PER, 8.9 WS, .149 WS/48, 1.7 BPM, 2.7 VORP
C, Patrick Ewing 1993-94: 37.6 MPG, 24.5 PPG, .496 FG%, 11.2 RPG, 2.3 APG, 2.7 BPG, 22.9 PER, 13.1 WS, .211 WS/48, 5.2 BPM, 5.5 VORP
We discussed the point guard rankings at length so we won’t revisit other than to remind that Ward was pretty good. At the two-guard, Starks always seemed to playing above his athletic ability while Houston and Sprewell seemed so fluid and easy. In the end, peak-Starks did more. He could pass and defend and, despite his weaknesses, was the best Knicks shooting guard (at peak). Ultimately, this team was Ewing and some guys who tried really hard but were only pretty good. Even Ewing wasn’t quite as good as his star center peers of that era.
These teams were a tribute to coaches who got players to buy in and play hard. Of course, this style of play could not work today. With zone defenses and the inability to hand check/arm bar, the nature of what is considered effective defense has changed. In addition, in the need for 3-and-D big men today seems funny when you see that the Knicks were shoehorning Smith, Mason, and LJ into small forward (two of whom rarely shot for more than 10 to 12 feet). But the distinction doesn’t matter. The key is to make a winning team within the context of the rules at the time. Riley and Van Gundy executed those plans quite well and are reminder to Knicks fans that there was a time when management actually had plans.