Shai Gilgeous-Alexander — Two-Way Profile (3 Seasons)
Source: BBall Index (bball-index.com) — written with the Basketball Analytics Expert Sage active.
The Question
Two questions, one frame: what is OKC actually getting from the offensive side of the league's most valuable player, and what does the team have to engineer around on the defensive side? The first question is easy to misread because the headline numbers are big in every direction; the second is easy to caricature in either direction (he's a MVP-runner-up defender / he's a turnstile). The honest answers are quieter than either.
Part 1: The Big Picture
LEBRON sits at 6.23 — the best individual mark in basketball this year and the highest among any OKC player. The three-year arc is 5.26 → 6.78 → 6.23, with the 2024-25 MVP season as the peak. The 0.55-point dip from 24-25 to 25-26 is inside the typical single-season noise band for top-of-league composites; the multi-year prior is "best-or-near-best player alive," not "declining."
The split is one of the cleanest in the NBA: O-LEBRON is doing essentially all of it. 4.34 → 5.91 → 5.43 — a top-1-or-2 offensive engine for three straight years. D-LEBRON is steady-mildly-positive (0.92 → 0.87 → 0.80), which for a 33% usage lead guard is closer to a feature than a flaw — most players carrying that offensive load post negative defensive impact, and SGA simply does not. Call it the right defensive read: not a stopper, not a leak.
LEBRON WAR at 11.96 in 25-26 (down from 14.58 in 24-25, partly because he's playing 33.2 minutes per game vs. 34.2 last year) is still the league lead among full-season players. Usage Rate has been pinned in a 32.8–34.8% band for three years — he's running OKC's offense at a rate that almost no one in basketball history has sustained efficiently. The fact that True Shooting % has actually risen (63.5% → 65.9%) inside that usage envelope is the unusual fact about his season.
| Metric | 2023-24 | 2024-25 | 2025-26 | Read |
|---|
| LEBRON | 5.26 | 6.78 | 6.23 | Top of league three years running |
| O-LEBRON | 4.34 | 5.91 | 5.43 | The engine, basically all of it |
| D-LEBRON | 0.92 | 0.87 | 0.80 | Mildly positive — feature, not flaw, at this usage |
| LEBRON WAR | 12.11 | 14.58 | 11.96 | League lead even with fewer minutes |
| Offensive Talent | 4.18 | 5.31 | 5.40 | Talent read still rising |
| Minutes Per Game | 34.0 | 34.2 | 33.2 | Modest load management |
| Usage Rate | 32.8% | 34.8% | 33.4% | Three years of historic sustained usage |
Part 2: The Offense
Efficiency at usage — the central fact
The right way to read True Shooting % at 65.9% is to put it next to Usage Rate at 33.4%. That's the rare upper-right quadrant of basketball — the volume × efficiency product is what makes him the league's best offensive player and what the model captures in O-LEBRON of 5.43. Effective FG% at 59.7% (up from 56.9%) and Overall Shot Making at 3.05 (up from 2.93) both confirm: he is making a higher rate of his own shots, on harder looks, than at any point in this three-year window.
The shot-quality side is the giveaway about how he plays. Overall Shot Quality at -0.69 is negative — meaning the shots he takes are below the league average in expected value. That's not a problem; that's the job. He is a self-creator who manufactures contested midrange and pull-up looks against set defenses, which by construction live below average shot quality. Overall Shooting Talent at 3.48 is the composite that prices that in: the volume-weighted blend of harder-than-average attempts × better-than-expected execution × creation kicker. Three-year arc 2.23 → 3.45 → 3.48 says the talent read keeps climbing even after the MVP season.
| Metric | 2023-24 | 2024-25 | 2025-26 | Read |
|---|
| True Shooting % | 63.5% | 63.4% | 65.9% | Career high inside high-usage envelope |
| Effective FG% | 56.7% | 56.9% | 59.7% | Same direction |
| Overall Shooting Talent | 2.23 | 3.45 | 3.48 | Composite still climbing |
| Overall Shot Making | 1.77 | 2.93 | 3.05 | Above-expected execution rising |
| Overall Shot Quality | -0.39 | -0.95 | -0.69 | Negative by design — self-created looks |
| Overall Shot Creation | 2.55 | 2.82 | 2.75 | Stable elite |
| True Usage | 53.8 | 57.7 | 51.7 | Slight dip with reduced minutes |
The self-creation evolution — fewer drives, harder iso, more midrange
The most interesting movement this year is what he's doing inside the offensive role, not how much he's doing. Drives Per 75 has dropped from 24.41 to 15.35 — a 37% cut in drive volume — while Rim Shot Attempts Per 75 is barely down (6.24 → 5.73) and Rim FG% is up to 71.6%. He's not driving less because defenses are stopping him; he's driving less because he's choosing midrange more, and converting at higher rates when he does drive. Finishing Talent at 3.72 (up from 2.68) is the talent read on that selectivity — better attempts, fewer contested ones.
The iso/midrange shift is where the talent climb actually shows up. Isolation Shooting Talent at 5.64, One on One Shooting Talent at 4.96, Self-Created Shot Making at 4.54 are all three-year highs; Midrange Talent at 3.99 is too. The pull-up game is the load-bearing skill: against a defense that walls off the rim, he gets to a contested midrange jumper that he hits at a stable elite rate. Floater Talent dropping from 2.58 to 0.37 is the only counter-direction reading — he's largely traded the floater for the midrange pull-up. The two are different shots from the same general area; the choice has shifted.
| Metric | 2023-24 | 2024-25 | 2025-26 | Read |
|---|
| Isolation Shooting Talent | 4.18 | 4.31 | 5.64 | Three-year high |
| One on One Shooting Talent | 3.10 | 3.55 | 4.96 | Same |
| Self-Created Shot Making | 2.96 | 4.34 | 4.54 | Same |
| Midrange Talent | 2.79 | 3.75 | 3.99 | Same |
| Floater Talent | 2.58 | 2.26 | 0.37 | Replaced by the midrange pull-up |
| Finishing Talent | 2.50 | 2.68 | 3.72 | Same — selectivity pays |
| Rim FG% | 69.6% | 70.2% | 71.6% | Same |
| Drives Per 75 | 24.41 | 21.60 | 15.35 | Choosing midrange over driving |
| Rim Shot Attempts Per 75 | 6.24 | 5.78 | 5.73 | Rim attempts steady |
| Drive Foul Drawn Rate | 9.4% | 9.4% | 9.6% | Foul-drawing rate intact |
| Contact Finish Rate | 27.2% | 22.3% | 23.2% | Steady inside the prior |
Playmaking and gravity — the downstream effects
Playmaking Talent at 2.53 is a three-year high, and Box Creation at 14.42 has climbed every year (11.86 → 13.86 → 14.42). The reading is that the ball-handling load isn't only producing his own shots — the assist-quality pipeline is widening too. Passing Creation Quality at 0.59 is back into clearly positive territory after dipping to -0.08 in 24-25.
The gravity numbers tell the team-impact story. On-Ball Gravity at 5.58 is the highest single-season figure for any guard the model rates, and the three-year climb (2.16 → 3.48 → 5.58) is the cleanest single trajectory in his profile. Defenses are bending toward him on every possession he handles the ball; that bend is what creates the open looks for his teammates that aren't measured in his own shooting line. Off-Ball Gravity at 1.30 is steady-positive but not the headline — defenders don't fear him as a spot-up threat, and OKC isn't asking him to be one.
| Metric | 2023-24 | 2024-25 | 2025-26 | Read |
|---|
| Playmaking Talent | 2.06 | 1.83 | 2.53 | Three-year high |
| Box Creation | 11.86 | 13.86 | 14.42 | Climbing each year |
| Passing Creation Quality | 0.93 | -0.08 | 0.59 | Back to positive |
| On-Ball Gravity | 2.16 | 3.48 | 5.58 | Cleanest single trajectory in his profile |
| Off-Ball Gravity | 1.12 | 1.37 | 1.30 | Steady-positive, not the headline |
Part 3: The Defense
A neutral defender, hidden by scheme
The right framing first: at 33% usage, the expectation is negative defensive impact. Most players who carry that offensive load post negative D-LEBRON, often by a lot. SGA's D-LEBRON of 0.80 — and the three-year band of 0.80 to 0.92 — is the read of a guard who is not a leak, which is roughly the best ROI you can plausibly extract from someone running an MVP-tier offense.
The on-ball metrics are honest: Perimeter Isolation Defense at -0.25 (three-year band -0.41 to 0.15) is mildly below average. He is not a stopper; he can be attacked one-on-one. Ball Screen Navigation at -0.27 (band -0.52 to -0.39) is similar — he can be hunted in pick-and-roll if a defense wants to. The honest read is that as an isolated defender he is somewhere between league-average and slightly below.
But the scheme reads tell you what OKC actually does with him. Matchup Difficulty at -1.06 means the players he's defending have easier offensive profiles than the league-average matchup — OKC is deliberately routing the harder assignments to Dort, Cason Wallace, and (situationally) Chet. % of Time Guarding Primary Ball Handlers has dropped from 23.6% (23-24) to 13.4% (25-26), and % of Time Guarding Shot Creators sits at 57.8% — most of his work is against shot creators who are not the opponent's primary handler. That's the system getting maximum mileage from his help-positioning ability without exposing him as the anchor of the perimeter defense.
| Metric | 2023-24 | 2024-25 | 2025-26 | Read |
|---|
| D-LEBRON | 0.92 | 0.87 | 0.80 | Mildly positive — not a leak at 33% usage |
| Perimeter Isolation Defense | 0.15 | -0.41 | -0.25 | Mildly below — can be attacked |
| Matchup Difficulty | -0.82 | -1.05 | -1.06 | Easier-than-league assignments by design |
| Ball Screen Navigation | -0.39 | -0.52 | -0.27 | Hunt-able in PnR |
| % Guarding Primary Ball Handlers | 23.6% | 13.6% | 13.4% | Moved off lead handlers |
| % Guarding Shot Creators | 72.4% | 49.9% | 57.8% | Still takes secondary scorers |
Help and activity — the quiet work
Help Defense Talent at -0.33 and Help Defensive Activity at -0.57 are both in the slight-negative band he's lived in for three years. The activity number is the more interesting read — he's not chasing rotations as much as a Cason Wallace or Dort would, which is consistent with role-conservation for the player carrying the offense. Steals Per 75 at 1.51 (down from 2.10) is in the band where stab-and-grab steals are part of the read but also subject to small-sample variance.
Defensive Role Versatility at 42.57 (down from 56.07) is the cleanest single-number read of "the role is narrowing." Combined with the % Guarding Primary Ball Handlers drop, this is a deliberate scheme decision: OKC is asking him to do less defensively, in exchange for more offense.
| Metric | 2023-24 | 2024-25 | 2025-26 | Read |
|---|
| Help Defense Talent | -0.64 | -0.41 | -0.33 | Inside a slight-negative band |
| Help Defensive Activity | -0.64 | -0.43 | -0.57 | Lower activity by design |
| Steals Per 75 | 2.10 | 1.82 | 1.51 | Inside small-sample band |
| Defensive Positional Versatility | 84.7 | 81.8 | 75.7 | Slight narrowing |
| Defensive Role Versatility | 56.1 | 51.9 | 42.6 | The clearest "role is narrowing" read |
Part 4: Team Fit — Strengths and Absorptions
What OKC gets
A top-of-league offensive engine running 33% usage at 65.9% True Shooting %, with On-Ball Gravity that bends defensive shape on every halfcourt possession. The cleanest way to put it: he is the player around whom a 16-point net rating starting lineup is constructed (per okc-starting-five-2025-26.md), and the iso/midrange creation he can execute against any matchup is the pressure release valve when the offense breaks down.
The 2025-26 evolution is the part the league should pay most attention to. He has cut his drive volume by more than a third while raising True Shooting %, Finishing Talent, Isolation Shooting Talent, and On-Ball Gravity. That's a stylistic shift toward midrange/iso supremacy without the wear-and-tear cost of the higher-drive 24-25 season. For OKC, that's both better playoff durability and a profile that's harder to scheme against — you can't load the paint to stop a player who's beating you with a contested mid-post pull-up.
What OKC absorbs
A guard-sized primary defender who can be attacked at the point of attack, with mildly negative isolation and ball-screen navigation reads. The team's answer is structural: Dort and Cason Wallace handle the toughest perimeter assignments, Chet's switching gives the scheme an emergency ceiling against guard creators, and Hartenstein's rim presence covers what gets through. The 1.06-point negative Matchup Difficulty says OKC is not pretending — they are routing the hard assignments away from him, deliberately and consistently.
The absorption isn't free. With % Guarding Primary Ball Handlers at 13.4%, the perimeter defensive load on Dort and Cason Wallace is correspondingly high. In playoff matchups against teams with two perimeter creators (Boston, Denver in a healthy year), the constraint is real — OKC has only so many minutes of Dort and Cason on the floor to handle two opposing creators while keeping SGA off them. The defense works in the regular season because the supporting cast is deep enough to absorb it; the playoff math is tighter.
The two-way verdict
He is the best offensive player in the league this year, by a margin that the impact composites are converging on rather than diverging from. The defensive side is correctly characterized as "not a leak, hidden by a deep scheme" rather than either of the two caricatures the discourse usually reaches for. The team-fit ROI is exceptional: an MVP offensive load that doesn't cost OKC defensive net rating, paired with a supporting cast designed precisely around what the player can and can't do on each end.