So far this season, the Rockets are 4-3 in games decided by five points or less. The fact that Houston has played only seven such games over a 50-game sample points to the great variance and variability in this young team’s performance. A 4-3 record doesn’t suggest that the Rockets are performing particularly poorly or exceptionally in close games. Nonetheless, I took a look at the available clutch statistics to see if I could find anything of interest, especially when compared with how the Rockets perform in general.

Unfortunately, I could not find any reliable team clutch statistics (if anyone is aware of a website that has such information, please let me know). As such, in lieu of a team-oriented analysis of clutch performance, I looked at individual performance for some of the Rockets’ most heavily-used players in the clutch. All statistics are courtesy of 82games.com, which defines the “clutch” as situations in the fourth quarter or overtime with less than 5 minutes remaining and neither team ahead by more than 5 points.

**James Harden:**Harden plays 90% of the Rockets’ available clutch minutes and is nothing less than an offensive machine—he is scoring 39 points per 48 minutes in the clutch while attempting 19 free throws per 48. Surprisingly, his effective field goal percentage (eFG%), which incorporates three pointers but not free-throws, is 38% in the clutch compared with 49% in all situations. Harden more than compensates for this poor shooting by attacking the rim more frequently (44% of his shots in the clutch are close to the basket, compared with 31% in all situations) and drawing fouls at a higher than usual rate. The upshot is that the Rockets are +6 points per 48 with Harden on the floor in the clutch and can rely on him as the focal point of their crunch-time offense.**Jeremy Lin:**Lin has played 70% of the Rockets’ clutch minutes while scoring 25 points per 48 minutes and registering a robust eFG% of 57%, an 11% improvement on his season average. Lin’s usage rate in the clutch is lower than Harden’s (he takes 12 shots per 48 minutes, compared with 26 for Harden) but he nearly doubles the rate at which he draws fouls despite the fact that 80% of his field goal attempts are categorized as jump shots. Compared with Toney Douglas, a player who has sometimes replaced Lin down the stretch in games (to much controversy and hand-wringing), Lin seems like an offensive maestro—Douglas sports a horrific 32% eFG% in the clutch, almost never draws any fouls, and functions exclusively as a spot-up shooter.**Omer Asik:**Asik enjoys 65% of the Rockets’ clutch minutes and has a net per-48 rating of +13 points. His scoring is atrocious (eFG% of 22%), but he blocks 12% of all shots on the floor and grabs a full 64% of all of Houston’s defensive rebounds in such minutes. Defensively, the Rockets give up a paltry 83 points per 48 minutes with Asik on the floor in the clutch.**Chandler Parsons**: Parsons is on the floor 86% of the time in the clutch, a rate second only to Harden’s. In such minutes, Parsons serves mainly as a spot-up shooter (84% of his field goal attempts are jump shots). Curiously, the Rockets’ offense is only scoring 96 points per 48 with Parsons on the floor, even though he attempts only 13 shots per 48 and his eFG% of 50% is only marginally lower than his season average of 54%.**Patrick Patterson:**Patterson garners 56% of the Rockets’ clutch minutes and sports a +15 net 48 rating despite the fact that his eFG% of 37% is well below his normal rate of 54%. Once again, Patterson is primarily a jump shooter in these situations, attempting two-thirds of his field goals via jumpers.**Carlos Delfino:**Delfino appears in exactly half of the Rockets’ clutch minutes and sports a respectable eFG% of 50% while taking jump shots on 92% of his field goal attempts.

A few observations follow from the above:

- In the clutch, the Rockets’ ball-handling stars (Harden and Lin) take an even greater share of the offensive burden and rely heavily on getting to the line to generate late-game offense.
- As usual, they are surrounded by a cohort of spot-up jump shooters in Patterson, Delfino, and Parsons, all of whom match or even exceed their typical rate of taking jump shots.
- The net-48 rating is all over the place and probably a relatively meaningless statistic in these limited sample sizes.

I’ll conclude with a few more caveats:

- Clutch statistics are notoriously unreliable and highly dependent on how “clutch” is defined. The necessarily limited sample size only complicates their interpretation and compromises their reliability.
- As with many NBA metrics, much of the available clutch statistics skew towards offensive measures and may not serve as a complete or accurate representation of a player’s defensive impact.
- In general, it may not even make much sense to focus on clutch statistics—points scored or allowed in the first quarter are just as valuable and important as those in the waning minutes of the fourth quarter or overtime.