Such a small sample size for that lineup data to be very useful, but still interesting nonetheless.

It's frustrating that our most efficient lineup only saw 15 possessions together.

Delly
Mills
Broekhoff
Anderson
Bogut

I was screaming out for this (or similar) lineup more often.

Spacing, spacing, spacing.

It seems like we never played all our good shooters at once: always subbing them in for each other. That's not how shooting works: if you want spacing they need to all be in at the same time.

Interesting how good Motum looks with this data. He completely restored my (lost) faith in him after that Bronze match. Played really hard on offense and looked like he genuinely gave a damn, even if he'll never figure out the defensive end.

If Bogut doesn't crumble to dust (a real possibility), slotting Ben Simmons into that Dave Anderson spot in that lineup in three years has got me excited.

Good point, Mick. My perception was that the lineups with a range of shooters seemed to do well. I might have a selection bias, though, as I've always liked having a stretch 4.

I was a bit surprised by the starting lineup as I didn't think Baynes would work so well as a 4 with Bogut. I recall Baynes being totally exploited defensively against opposition stretch 4s, and the perimeter options on offense with a Baynes-Bogut combo were limited.

So I was curious to see how lineups featuring Baynes and Bogut playing simultaneously fared. I removed the China and Venezuela games because I was more concerned about tough opponents (besides, Bogut was rested versus China, so that would skew the data). The starting lineup combination with Baynes-Bogut was just +4 in 92 possessions (+4.4/100), but in total, Baynes-Bogut lineups were +13 from 123 (+10.6/100), compared with non-Baynes-Bogut combo lineups +17 from 307 (+5.5/100). So I suppose my concerns were not backed up by this data.

Again removing the China and Venezuela games, I was also curious to see the on-off +/- numbers. It appears the Spain game hasn't been updated into the website yet.

Bogut
ON: +34 from 238 (+14.3/100)
OFF: -4 from 192 (-2.1/100)
ON-OFF difference: +16.4

Delly
ON: +45 from 258 (+17.4/100)
OFF: -15 from 172 (-8.7/100)
ON-OFF difference: +26.2

Mills
ON: +7 from 338 (+2.1/100)
OFF: +23 from 92 (+25/100)
ON-OFF difference: -22.9

Ingles
ON: +5 from 297 (+1.7/100) * (+5.1/100 without the Broekhoff at PF experiments)
OFF: +25 from 133 (18.8/100)
ON-OFF difference: -17.1

Baynes
ON: +29 from 244 (+11.9/100)
OFF: +1 from 186 (+0.5/100)
ON-OFF difference: +15.9

Andersen
ON: 19 from 204 (+9.3/100)
OFF: 11 from 226 (+4.9/100)
ON-OFF difference: +4.4

Martin
ON: +14 from 69 (+20.3/100)
OFF: +16 from 361 (+4.4/100)
ON-OFF difference: +15.9

Broekhoff
ON: +15 from 153 (+9.8/100)
OFF: +15 from 277 (+5.4/100)
ON-OFF difference: +4.4

Bairstow
ON: -3 from 88 (-3.4/100)
OFF: 33 from 342 (+9.6/100)
ON-OFF difference: -13.1

Lisch
ON: -13 from 138 (-9.4/100)
OFF: +43 from 292 (+14.7/100)
ON-OFF difference: -24.1

Motum
ON: -8 per 48 (-16.7/100)
OFF: 38 from 382 (+9.9/100)
ON-OFF difference: -26.6

Goulding
ON: +10 from 52 (+19.2/100)
OFF: +20 from 378 (+5.3/100)
ON-OFF difference: +13.9

OVERALL: +30 from 430 (+7.0/100)

It makes a big statement about the irreplaceable importance of Bogut and Delly. They both change the dynamics of the offense immensely while also providing great defense.

Visual metrics:
- Lisch and Martin's high turnover% (highlights the lack of ball handlers off the bench)
- Ingles low rebound% (significantly lower than Broekhoff's)
- Motum O-board% (trademark sneaky putbacks?)
- Bairstow D-board% (a huge 28%)
- Top O Rtg's of 140+ on the team include Delly and Bogut... and Martin! (how was that possible? Incidentally, Martin's usage% was just 5.7%)
- Martin's 89.9 defensive rating led the team
- Bairstow's offensive rating of 89 was clearly the worst on the team, while his assist% of 2% was clearly the lowest (struggles executing?)

(All data again excluding China/Venezuela games)

Some 2-man combos :

Mills + Lisch = -27 from 70 possessions (-38.6/100)

Delly + Lisch = +11/41 (+26.8/100)

Delly + Mills = +32/192 (+16.7/100)

Martin + Mills = +10/58 (+17.2/100)

Martin + Delly = +2/3


Based on this data, it appears Lisch really struggled combining with Mills. This fits with my recollection of Lisch really struggling to initiate and run the offense. Once combined with Delly, Lisch lineups work better. Which made me wonder how Mills-lineups fared without Delly as floor general:

Mills on with Delly: +40/239 (+16.7/100)
Mills on without Delly: -30/85 (-35.3/100)

Without Delly as floor general, Mills-lineups struggle. This consolidates the viewpoint that we have held for a long time: Mills is not a PG. While Mills can put up points as a PG, but I'd argue that his team performs much better when he plays SG. The figures of Mills without Delly could be the result of the bench replacements dragging the +/- down, so let's do the same for Delly to see the contrast, to see how Delly fares with benchies.

Delly on with Mills: +40/239 (+16.7/100)
Delly on without Mills: +16/63 (+25.3/100)

No issues for Delly there. Delly is clearly the best floor general we have. Without him, the offense struggles. Delly's offensive rating was a blistering 149 (which was better than any main rotation team USA player, by the way).

Some other 2-man lineups:

Andersen + Bogut: +12/86 (+14/100)

Andersen + Baynes: +22/81 (+27.2/100)

Andersen + Bairstow: -8/25 (-32.0/100)

Curious to see Andersen actually seem to combine better with Baynes.


Isolating players at a position:

Small forward:

Ingles at SF: +13/261 (+5.0/100)
Broekhoff at SF: +24/117 (+20.5/100)

Although Broekhoff's shot was a bit off, Broekhoff was actually looking for his shot. When he is shooting well, the team looks so much more dangerous. Broekhoff plays good D and competently executes the offense, so I'm not so surprised by his effect on the team. Although I can see the benefit of Broekhoff providing firepower off the bench, the most potent backcourt lineup (Delly-Mills-Broekhoff) seems to have been seldom used:

Delly-Mills-Broekhoff: +20/49 (+40.8/100)

Delly-Mills-Ingles: +5/135 (+3.7/100)

My perception was that the Delly-Mills-Broekhoff combination seemed the most potent combination in 2014, too.

Power forward:

Baynes at PF: +12/101 (+11.9/100)
Andersen: +24/201 (+11.9/100)
Motum: -2/36 (-5.6/100)
Bairstow: +3/63 (+4.8/100)

Re: Motum, I believe this data doesn't include the Spain game.

Centre:
Bogut at C: +32/238 (+13.4/100)
Bogut at C (without Baynes on): +20/137 (+14.6/100)
No Bogut at C: -4/192 (-2.1/100)
Baynes at C: +17/143 (+11.9/100)
No Bogut/Baynes: -21/49 (-42.9/100)
Andersen at C: -15/37 (-40.5/100) *(note: only played C with Bairstow, Motum and Broekhoff at PF)

Lineups with either Bogut or Baynes at C were respectable. Any lineup without either Bogut or Baynes at C really struggled. Not surprising. Although that might have been when Lemanis was desperately experimenting with lineup combinations in the Serbia SF.

Hi RJD,

You're correct, the lineup data doesn't include the bronze medal game against Spain.

Thanks for all the feedback too guys, great to see people taking an interest in this stuff.

What really seems to come out of all of this data is not just that Delly-Mills-Broekhoff was our best option in the back-court (and under-utilised), or that Bogut or Baynes needed to man the middle (pretty obivous), or even that Andersen was incredibly valuable off the bench (has he ever not been, for the Boomers?)... but that Ingles really wasn't much use at all.

Maybe the Newley decision was as bad as a lot of people thought it was at the time, not just because he should have been out third option at 3, as 10th or 11th man, but because he would have been more effective than Jingles as the first or second option at 3.

I still think he should have been taken over Goulding. The international game is so much more demanding of taller wing players, and I don't think Goulding is being looked at as a long term option for Boomers representation.

"Lisch and Martin's high turnover% (highlights the lack of ball handlers off the bench)"

It actually just shows the limitations of usage stats. Martin only had three turnovers in 69 minutes.

"Unfortunately the randoms of stats mean nothing when they don't play together. All theoretical unfortunately"

I'm not sure you understand...that's the whole point of lineup data...it shows you when they play together...

Thanks, Paul. I'm pretty sure I misinterpreted this one. How is turnover% calculated?

Martin had 3 turnovers in 69 minutes, and Lisch 5 turnovers in 112 minutes (as a comparison, Delly had 11 turnovers in 181 minutes). Can't really complain about those stats!



@andrew, keep up the awesome work.

100 * TOV / (FGA + 0.44 * FTA + TOV)

Not shooting the ball gets you a high turnover percentage.

Thanks koberulz, that explains it.


This fits in with team seemingly transforming immediately after Bogut came back:

Delly+Bogut = +33/164 (+20.1/100)
Delly+(noBogut) = +7/75 (+9.3/100)

Assuming I did it right, Delly & Lisch & Broekhoff together was an impressive +54.

That's three quality defenders and unselfish players to boot.