Numbers hold no prejudices, preconceived notions. Not colored by experience, personal views, agendas, or the ever-revered ‘eye test,’ they simply reflect what’s happening.

To be fair, though, numbers, and, as it’ll relate specifically to this post, statistics, don’t tell you the story. They give you the end, the conclusion, the finale, while tearing out 95% of the pages. They’ll give glimpses of chapter names, characters, and just enough of the narrative arc to let those who use them craft their own version of how the numbers got there (guilty, so guilty). Looking at the tables and numbers in this post, or more advanced statistics generated by people with more time, data, and skill/smarts than the author, will never tell you more about the game than watching (a lot), coaching, or playing. They’re just numbers; not basketball.

What they can be used for is refining observation and understanding of the game and players within it, as well as debunking and/or uncovering explanations for why a team and its players play how they do. They can point out, despite a running narrative that Villanova’s offense runs better with Ryan Arcidiacono at the point, that the team scores 14 points more per 100 possessions when he sits, as compared to when he’s in the game. They can’t note the effect he’s had on team chemistry with his effort, grit, and late game heroics, and whether that effect has been greater on the team than his more-than-occasionally negative weight on the offense.

Beyond the simple things stats can’t account for, the actual mathematic error inherent in these statistics, especially drawn from a small (though growing) sample size, should be mentioned. Statistical noise, especially in many of the tiny samples the larger (read: 4- and 5-man) lineups have logged, is common. How a certain group has played for 40 disconnected minutes over half a season doesn’t write the book on how they would play in 40 minutes of continuous game time, or 40 more non-linear minutes the rest of the season, and there are often stats skewed by the tiny samples. These particular stats don’t account for game situation, score differential, or the strength of the opponent/players in the game – you’re basically looking at back-of-the-napkin scratch, as far as real advanced stats go. There’s more error in non-adjusted stats, and I would love to be able to account for the strength of the other team’s lineups on the floor, and incorporate stats generated from film study – but it’s difficult, time-consuming, I don’t have a DVR, and though I actually love writing/generating these for me and the 5 other people on this site who both enjoy and understand advanced stats, that shit’s hard, and I’ve already got a full-time job.

With all that said, tracking lineup stats, even as I do it, is an excellent way to look at how players/groups of players play together, what has and hasn’t worked, and what might in the future. So let’s take a look at how the team has played, in all of its iterations, through the 18th game of the season.

Standard intro used for these particular posts:

Please see the introduction of the first post for the methodology behind the analysis, and explanation of many of the stats referenced in the tables and the article.

These stats are taken from play-by-play data available from StatSheet, slightly modified to correct substitution patterns (they generally don’t note who goes in/out at halftime). The Lafayette game was not included, as I couldn’t ‘fix’ the substitution patterns.

The 9 players who have seen actual rotation minutes - Ryan Arcidiacono, Darrun Hilliard, James Bell, JayVaughn Pinkston, Daniel Ochefu, Josh Hart, Tony Chennault, Dylan Ennis, and Kris Jenkins – and all combinations of them in 1-, 3-, 4-, and 5-man groups – are in this analysis.

‘Remember – these statistics do not indicate how well the individual player (or group of players) does statistically, but how well the team plays around him/them while he’s/they're in the lineup. Obviously, the player’s stats are included if he’s in the game, but [this is] the overall picture of how well the team plays with (or without) a certain player, or group of players.’

I’m only going to include the slightly modified Four Factors sheet in the article itself – all stats and numbers referenced but not shown exist in my spreadsheets, and interest me, but probably not most of you. Will gladly send/post tables if anyone’s interested in the nuttier and boltier portions of the stats, just let me know.

Let’s start with the with and without 1-man lineups, showing how the team has played around the player when he’s on the floor (1st table), and how they play when he’s out (2nd table).

‘With’ 1-man Lineups

‘Without’ 1-man Lineups

By most measures in these two charts, Ryan Arcidiacono has the most negative effect on the team. His lineups shoot the 2nd worst, rebound the 3rd lowest percentage of their own misses, score the least points per 100 possessions (tied with JayVaughn Pinkston’s groups at 112), and rank in the middle of the road in TO% and free throw rate. Correspondingly, lineups that don’t feature Arch rank tops in offense, rebounding rate, and eFG% - though, as a bone for his supporters, the TO% when he’s not in the game is one of the highest – support for his calm, measured running of the offense. But, a 1 percent difference in TO% doesn’t an offense run, and shooting, rebounding, and free throw generation jump significantly when he sits. He also helps crater the 3 point shooting – lineups without Arch shoot 47% - FORTY SEVEN PERCENT, 3 HATERS – in the 180 minutes he HASN’T been on the floor (44 of 93). He’s singlehandedly making Villanova an average 3 point shooting team.

My issues with Arch have been well-documented in these posts. However, there’s been a noticeable recent improvement in his offensive game – he’s been shooting out of his mind on long 2 point shots – almost 64%! - and he’s reining in the rampant inefficient chucking that has marred his game, albeit in fits and starts. And, ignoring the Creighton game (it’s less painful that way), has been getting beat off the dribble much less. The shooting on long 2 pointers is likely unsustainable, though – he only shot 38% on similar jumpers last year, and a 25 percentage point jump is near impossible over the course of a season –and he’s still prone to launching ill-advised 3s off the dribble. There’s no denying Arch is an important cog on this team, and he’s going to get a ton of minutes – but there’s pretty clearly a lot he can improve on. I’m going to keep rooting for a slight minutes reduction, and more minutes off the ball – he seems to be a much better shooter/shot decision maker when he’s not running the point. Incremental improvement in his game would be huge for someone who logs such a heavy minutes load. 50% 3 point shooting team, you’ll only live in my dreams.

Oh, and as a last nod to his problems on defense, lineups that don’t include Arch let opponents shoot 23% from 3 point range. This has to be influenced by late 3 point chucking when Arch is pulled towards the end of blowouts, but it’s still an insanely low number.

Big props to the Big Chief for boosting his lineup’s O Rtg with some destructive offensive performances in the 5 games or so since I last updated this data. He’s been hanging near the bottom for most of the season, mostly due to a bunch of turnovers and playing with the starters (who generally have lower offensive ratings and better defensive ratings than the reserve lineups). His recent performance has been enough to boost his lineups to the top of the starters, with those including him getting 116 points / 100 possessions. Not a huge difference between him and the rest of the starters, but the big difference is how much he improved his lineup rating in just 5 games. Truly impressive – along with the sky-high offensive rebounding rates (1st) and eFG% (3rd or 4th) his lineups sport.

Mildly disturbingly, he’s still barely made an effect on the defensive end – lineups without him are actually allowing 0.05 points per possession less, in more minutes (328 on vs 357 off), than those with him. This is actually trending negatively, as it’s been at a dead heat for most of the season. More troubling is the fact lineups without him are allowing opponents to shoot 46.8% at the rim – best (lowest) of the ‘without’ 1-man lineups, and lineups with him – correspondingly, allow a high FG% at the rim – 53+%. Basically, Ochefu seems to do a poor job of altering attempts at the rim that he DOESN’T block – though, to be completely fair to his abilities, he is an excellent shot blocker (blocks nearly 10% of opponent shots when he’s on the floor). Certainly something that would be nice to see him improve on as the year progresses, in terms of his positioning and rim protection when he can’t get up for the block.

A couple quick shots, before I get carried away with the most general of these lineup analyses:

The reserves we have are pretty clearly better than the ones they have. Tony C, Hart, Ennis, and Jenkins continue to light up the scoreboard when they’re on the floor, clocking in with the 4 highest lineup-associated offensive ratings.

A quick nod to Tony C, the statistical darling of my earlier FanPosts – he’s still helping the team kill it on offense, but his defensive impact has pretty clearly slipped lately, both in the stats and on-tape. He can’t stay in front of his man, and isn’t generating steals at the same rate he was earlier in the season – and his lineups are letting opponents pour it in. Still, his lineups score 120 points / 100 possessions, highest of the point guards, and shoot 43.75% from 3. Their propensity for getting beat off the dribble aside, Tony C profiles as the anti-Arch – hates shooting, drives THEN passes, and helps the team shoot better from 3. Man, I wish he played (sustained - it was so much better early in the year) better defense, or could shoot. Is Phil Booth here yet?

Ennis and Hart are the best perimeter defenders, per these 1-man lineups, at defending the 3. Lineups including them limit opponents to under 30% of their attempts from the perimeter. Preventing 3 attempts is always more important than the percentage a team hits. Teams actually have control over the number a team takes, not as much with the percentage that go in.

With that, we’re going to skip right past the 2-man lineups – done in the last post, and not a lot of new ground to cover there – and launch, for the first time, into the best 3-man lineups the team has.

I’m personally a pretty big fan of these lineup groups, as far as analysis. There are 84 different pairings of the 9 major rotation players, and nearly all of them have logged NON-meaningless amounts of minutes. 66 of those 84 combos have logged at least a full game worth of minutes (40+, for clarity), and all but one has hit more than 10. May not seem like much, but you’ll see what I’m talking about a bit further down. This list – biggest chart, promise – can be seen below, along with their performance in the Four Factors, O and D Rtg, and point differential. The chart is ordered by minutes logged.

Only considering lineups with more than 40 minutes logged…

3-Man Lineups (40+ Minutes)

Let’s run down the leaders in each category real quick. The most used lineups all include some combination of Arch, Hilliard, Bell, and Pinkston, and have all logged 245+ minutes on the floor.

The best shooting lineups (with meaningful minutes in the tank) are nearly all composed of bench players and one starter, or all bench players. The top 10 is as follows:

3-Man (eFG%)

Our bench is good, guys.

The best shooting lineup with 120+ minutes logged together is James Bell, JVP, and Chiefy, clocking in at 58.42% - which, incidentally, is also the first lineup encountered containing ALL starters (17th overall, 15th among lineups with 40+ minutes logged).

The TO% standings (top 10, as it will be from here on) are as follows:

Note – generally, I will be eliminating lineups that log less than 40 minutes from these standings, but I left the triple PG lineup in to show the one good thing it does – take care of the ball. Sorry, hate those lineups, including the two PG. It sacrifices size advantage at the 2 and 3, puts two bad perimeter defenders together, can’t rebound, and just generally sucks. NO MORE, JAY.

3-Man (TO %)

Arch is in almost all of these lineups – he can protect the ball well, that’s for sure. It’s mildly surprising to see Ochefu in the group at the top, simply because of his problems early in the season, but it’s not upsetting. Not a ton of analysis to provide, for most of these – it’s more about the info than parsing out why a specific group of 3 has performed well in its minutes together. Fitting a narrative to them is tough and would generally be outright speculation.

Offensive Rebounding! (Same Process)

3-Man (O Reb%)

As you’d expect, at least one of Josh Hart, Daniel Ochefu, and James Bell are in most of these lineups – not a huge leap to assume you’d find the best individual offensive rebounders on the team in these rankings. The Jenkins-JVP-Arch combo is pretty surprising, but certainly encouraging as far as the value of a JVP-Jenkins frontcourt for portions of the game. Dylan Ennis also appears a bunch, as one of the better rebounding guards we have. Love the 40% rebound rate logged for the starter group of Arch-JVP-Chief in 156 minutes.

And, the fourth factor – free throw rate.

3-Man (FTR)

Most of the samples are relatively small, and I feel FTR is one of the advanced stats most easily skewed, especially by groups present at the end of blowout games (see the TC-Ennis-Jenkins grouping atop the rankings), but you have to love Hilliard-Bell-JVP shooting more than a free throw per two shots. JVP is present in most of the higher-minutes-logged-but-still-in-the-top-10 groups - no surprise.

And then, all in a row, best offensive, defensive, and overall differential groups.

3-Man (O RtG)

3-Man (D Rtg)

3-Man (Differential between O Rtg and D Rtg)

As should probably be expected by now, the top 10 in offensive rating is populated by Villanova’s reserves. They are so much better than what most teams have coming off the bench, and that depth is arguably the biggest difference between last year’s team and this one. Deep bench scoring and more energy have keyed an unappreciated portion of the team’s 16-2 run. It should be noted, though, that not all their minutes come against the other team’s bench – most other teams barely have one, anyway. They perform well against everyone.

The best all starter group can be found WELL down the list, with James Bell, JVP, and Big Chief scoring 120 points/possession (ranked 31st of the 84 groups) in 130 odd minutes logged.

I left a few more groups on the defensive table to let the 2 people who read this far appreciate the defensive strength starter groups are showing – in 200+ minutes logged, the Hilliard-Bell-JVP and Arch-Hilliard-JVP lineups are holding teams to 85 points scored / 100 possessions – excellent defense, especially by starters, over such a sustained stretch. Gotta love Hart cropping up in one of the top groups – of all perimeter defenders – in Darrun, James Bell, and Hart – helping hold lineups they face to 80 points / 100, especially after some early season struggles were noted for him. Wish Ochefu were on here more, BUT his absence doesn’t necessarily speak to problems for him. Just want him to show his importance on defense.

Again, left a few extra names on here – differential is difference between O Rtg and D Rtg - to show some Ochefu, and more all-starter lineups, near the top despite much higher minutes logged than most of the other leaders. Tajh is the star of this list, popping up on NINE of the top 14 differential-leading lineups – the man’s just a better player than everyone thinks, even during an extended cold streak. So important to this team.

And, some quick hitters on some stats not shown so far:

The lineup of Josh Hart, Tony Chennault, and Dylan Ennis – in almost 80 minutes – is hitting OVER 60 PERCENT of its 3s. Nothing gets me going more than raining 3s (don’t make it sexual, guys), and these guys are head-and-shoulders above the next best group (53.33%...HA!), hitting 28 of 46 3 point attempts.

In 104 minutes, Arch-Ochefu-Hart is dishing – an A/T ratio of 2.14, with 65% of the baskets being assisted, and taking away, with a steal percentage of 15.79% - generating a steal on almost 16% of their opponents’ possessions, and causing turnovers on over a quarter of them (26.02%).

In 80 minutes on the floor, Arch-Hilliard-Ennis and the other 2 interchangeable parts have blocked an astonishingly high 16.25% of opponent’s shots. Have to think it’s more due to whoever they’ve been paired with, but still a high number for an appreciable sample.

And, on to the 4- and 5-man lineups. While I consider these valuable, there are very few lineup combinations that stick together for long. Outside of his starting lineup, Jay doesn’t appear committed to have any group of 4 or more play together – combined, the 4- and 5-man lineups have just 19 combinations (17 4-man, 2 (TWO!) 5-man groups) who have played a full game together (again, 40+ minutes). Won’t bother with ranking these, or presenting too much individual comment on the groupings.

4-Man Lineups (40+ minutes)

The combo of Arch, Darrun, James Bell, and Chief is shooting at an eFG% clip of almost 60% - an excellent, excellent mark over 162 minutes. Helps put it among the best offensive groups of the highest-usage pairings. That group, Arch-Bell-JVP-Chief and Darrun-Bell-JVP-Chief are all scoring more than 118 points / 100 possessions. Those latter two groups? Also pulling down more than 45% of the team’s misses. Other than highlighting those few stats (and the excellent defense played by almost all the starter-heavy groups), you can pretty easily see everything else I would talk about. There are only 17 groups on the list.

Quick Hits

You know I had to go back to the 3 point shooting well. Despite just clearing the completely, utterly arbitrary 40 minute threshold, JVP-Hart-Chennault-Ennis (recognize those last 3 names????) are hitting SEVENTY PERCENT of their 3s. It’s much higher than their 2 point percentage of 55%. More than a little nuts. That combo, with an O Rtg of 1.43/possession, would score 96 points in an average college game (67 possessions) if they strung that kind of shooting over an entire game. Found the Creighton antidote, guys.

Hilliard-Bell-JVP-Hart holding the opposition to 37.51% eFG% in 52.10 minutes – excellent. Is it possible JVP’s a better defensive center than Ochefu? I must be getting tired.

But wait!!! Take a look at the only two 5-man lineups with over 40 minutes (crazily, there are only 22 lineups with more than 10 minutes together).

5 Man Lineups (40+ minutes)

When you substitute Josh Hart for Danny Ochefu, presumably moving JVP to center, the defense becomes incredibly stingy, allowing only 69 points / 100 possessions over those 46 minutes, compared to 89 points / 100 in the 110 minutes Ochefu has played with the same lineup. I’m chalking this one up to more of a sample size thing than any indication JVP’s been playing better defense at center, but it’s at least interesting to look at.

That’s the end of another long one, and thanks for staying with me. I pretty much fell asleep writing it, and I care about it more than you do. If you read this and want to see more similarly-stats-based articles, feel free to email me with or post some ideas in the comments. Don’t want to only keep writing out these lineup analysis articles, but don’t have any other interesting ideas at the moment.