The Turnaround Playbook 

What the Data Says About Which Rebuilds Actually Work 

The Buffalo case

Twenty games into a recent season, our model projected Buffalo to finish with 74 points. They finished with 109 — a 35-point miss, and one of the largest in our five-season dataset. The points didn't show what was coming. The underlying shot-attempt data did.

That gap is the starting point for this case study. Not every rebuild is the same rebuild, and the standings are often the last place to see which kind you're in.

The archetype matrix

Every rebuilding organization can be located on two axes: CF% — Corsi For percentage, the share of all 5-on-5 shot attempts a team generates out of the total attempts in the game, used here as a proxy for underlying system strength — and points trajectory, the direction a team's results have moved over recent seasons. Plotting these two dimensions against each other produces four quadrants, and which quadrant a rebuilding team sits in predicts what the next three seasons are likely to look like far better than its current record does.

A team can be rising in points while its CF% stays flat — a sign the improvement is being driven by percentages (shooting luck, save luck) rather than a genuinely improving system. A team can also look unremarkable in the standings while its CF% climbs steadily underneath — the exact pattern that preceded Buffalo's jump.

The Turnaround Story teams

Four organizations in our five-season dataset occupy the quadrant we call Turnaround Story: low CF% relative to the league, but a clearly rising trajectory in both points and possession quality. This is the group worth watching most closely, because it's also the group most likely to be misjudged by anyone looking only at the standings.

Each of these organizations is at a different stage of the same underlying process. Mapped against a well-documented business transformation — IBM's nine-year turnaround under Lou Gerstner in the 1990s — the pattern is strikingly similar. Gerstner's first two years were pure stabilization: stop the bleeding, resist the pressure to break up the company, don't chase a quick fix. Years three through five were system installation — building a new operating identity and sticking to it long enough for it to take hold. Years six through nine were compounding returns, where the new system's results built on themselves and became difficult for competitors to catch.

Some of the Turnaround Story teams are in an earlier phase of that arc — still stabilizing, with the underlying CF% just beginning to move. Others are further along, with two or three years of consistent possession gains already banked, closer to the point where the system's compounding should start to show up decisively in the standings. The point isn't to predict exact timing — it's that "how long has this team been building, and how consistent has the underlying trend been" matters more than "what's their record right now."

The critical warning: the panic trade

The single most common rebuild failure in the data isn't organizational incompetence. It's a well-intentioned overreaction — the trade made right before the breakthrough, in response to a disappointing season that the underlying CF% trend said was still on track.

This is the exact pattern IBM's board pushed against in year three of Gerstner's tenure, when frustrated directors wanted to break the company apart rather than wait out a transformation that was, by the numbers, working. Gerstner's response was to hold the course. The organizations in our dataset that broke through were, without exception, the ones that didn't blow up a rising CF% trend chasing a faster fix. The ones that traded away futures for immediate relief interrupted a trajectory that the data suggests would have paid off within a season or two.

For a front office evaluating its own rebuild, the mistake to guard against isn't complacency — it's impatience dressed up as decisiveness.

Five questions for a GM in a rebuild

1. Has your CF% moved in a consistent direction over the last three seasons, independent of your record?

2. When your points have improved recently, was that improvement accompanied by a genuine possession gain, or mostly by percentages (shooting, save performance) that are less likely to repeat?

3. How many seasons has your current system — coaching staff, identity, core roster philosophy — actually been in place?

4. If you removed this season's record entirely, would the underlying shot-attempt trend still tell a story of progress?

5. Is the next proposed move an investment in the system already in motion, or a departure from it made under pressure?

Teams that can answer these five honestly have a real read on whether they're building toward a breakthrough or simply calling a plateau a rebuild. The diagnostic is only the first step — knowing where a specific rebuild sits on this matrix, and what stage of the transformation arc it's in, is where a direct engagement picks up.

© 2026 ORRO AI Genius, LLC. All rights reserved.