Brian Oliver
1990–91 to 1997–98 · 4 seasons
No. 2522 Career Impact Rank
282.3933 Career Composite Lower is better
NBA player comparison
Career comparison · Regular season
1990–91 to 1997–98 · 4 seasons
No. 2522 Career Impact Rank
282.3933 Career Composite Lower is better
1968–69 to 1969–70 · 2 seasons
No. 1309 Career Impact Rank
130.3133 Career Composite Lower is better
Joe Kennedy's best-3 average finish (128.6667 vs 278.6667) and best-5 average finish (131.2 vs 284.4) and best-10 average finish (133.1 vs 288.7) gives the lower Composite after weighting. The Career Composite gap is 152.08; lower is better. A model-defined difference, not a confidence interval or proof of a basketball advantage.
The career ranking includes unfilled season slots for Brian Oliver and Joe Kennedy. The calculation below shows the published replacement rule; it does not predict how a career will finish.
| Player | Best 350% weight | Best 530% weight | Best 1020% weight |
|---|---|---|---|
| Brian Oliver | 278.67 | 284.40 | 288.70 |
| Joe Kennedy | 128.67 | 131.20 | 133.10 |
These three windows determine Career Impact Rank. See the exact calculation →
The same ranking years, viewed through their basketball statistics. Descriptive seasonal averages, not additional ranking inputs.
Numbers show the stat. Bars show league standing. Average season percentile compares each selected year with eligible NBA players in that same season, then averages those standings using the selected weights. It is not a percentile among careers or extra ranking points. Gray bars describe role or workload, not quality.
Green + ▲ Leads marks the better displayed stat, not the percentile. Equal displayed values, missing comparisons and workload rows stay neutral. These row-by-row leads are not extra ranking points or proof of a meaningful basketball difference.
All bars share a 0–100 scale, with a midpoint at 50. Higher means a higher within-season standing; turnover avoidance reverses the direction so higher means fewer turnovers. The averages do not measure differences in era strength. Raw values are weighted seasonal averages, not pooled career stats.
Each season uses all NBA players with at least 450 minutes and an available statistic. Shooting measures require relevant attempts; lineup estimates require at least 1,000 possessions and an eligible sample. Cohort sizes vary by metric and season. Annual midrank percentiles retain full precision before averaging. Every required season must have both a valid statistic and a valid comparison pool. Impact Rate is a published model score, not a second percentile.
Seasons are selected by published league finish, with the earlier year breaking ties. Best 3, 5 and 10 overlap intentionally. Missing years are never replaced with invented statistics.
These players do not yet share a complete Best 3 ranking window. A side-by-side statistical profile is unavailable. You can still inspect available observations below; missing years will not be filled.
Regular season only. Missing data are not zero. Every statistic uses the same selected years. Complete shorter windows may be available when the full profile is not.
Select a complete common ranking window and let its source comparisons finish loading to preview a graphic.
Career Composite weights average league finishes over the best three, five and ten regular seasons at 50%, 30% and 20%. Lower is better. It is not a cumulative box-score ranking.
Green marks the better result for a directional measure. Playing time describes workload, not quality. Missing data are not zero.
Joe Kennedy's better best-3 and best-5 and best-10 league finishes produce the lower Composite after the 50% / 30% / 20% weighting. Peak impact and total career value are separate evidence, not extra points in that formula.
Joe Kennedy's advantages contribute 152.08 Composite points, offset by 0 in the other direction. The net difference is 152.08. Displayed values are rounded; the calculation retains full precision.
Regular season only. Impact figures are estimates; small gaps are not exact. Playoffs, awards and titles do not determine the verdict.
Games: 118, compared with 86. APG: 1.0, compared with 0.8. TS%: 43.6%, compared with 42.8%. These are distinct measures, not votes or additional points added to the ranking.
Points-above-replacement rank: #3782, compared with #4380. Peak era-centered impact: -2.6, compared with -3.6. Points above replacement: -35 pts, compared with -83 pts. These are distinct measures, not votes or additional points added to the ranking.
Only the three league-finish windows below determine Career Impact Rank. Career Value and peak impact are separate measures; neither adds points to this calculation. Lower composite scores rank higher.
282.393 Composite · Lower is better · No. 2,522 all-time
| Window and finishes used | Average finish Lower is better | Weight | Contribution Lower is better |
|---|---|---|---|
| Best 3No. 2501990–91Recorded box inputsIncludes bounded historical defense estimate2 open slots × No. 293 | 278.667 | 50% | 139.333 |
| Best 5No. 2501990–91Recorded box inputsIncludes bounded historical defense estimate4 open slots × No. 293 | 284.400 | 30% | 85.320 |
| Best 10No. 2501990–91Recorded box inputsIncludes bounded historical defense estimate9 open slots × No. 293 | 288.700 | 20% | 57.740 |
139.333 + 85.320 + 57.740= 282.393
Open-season rule: every unfilled slot uses No. 293, this player's stored replacement finish.
130.313 Composite · Lower is better · No. 1,309 all-time
| Window and finishes used | Average finish Lower is better | Weight | Contribution Lower is better |
|---|---|---|---|
| Best 3No. 1161968–69Reconstructed historical estimateIncludes bounded historical defense estimate2 open slots × No. 135 | 128.667 | 50% | 64.333 |
| Best 5No. 1161968–69Reconstructed historical estimateIncludes bounded historical defense estimate4 open slots × No. 135 | 131.200 | 30% | 39.360 |
| Best 10No. 1161968–69Reconstructed historical estimateIncludes bounded historical defense estimate9 open slots × No. 135 | 133.100 | 20% | 26.620 |
64.333 + 39.360 + 26.620= 130.313
Open-season rule: every unfilled slot uses No. 135, this player's stored replacement finish.
Averages, contributions, and totals are displayed to three decimal places. The ranking calculation keeps full source precision. Read the full method →
These are each player's averages across up to ten best seasons. Career Impact Rank gives the Best 3, Best 5, and Best 10 average league finishes 50%, 30%, and 20% window weights, giving the most weight to the highest peak. The model accounts for position, role, pace, playing time, and team results.
Brian Oliver: 1 qualifying seasons. Joe Kennedy: 1 qualifying seasons.
| Metric | Brian Oliver | Joe Kennedy |
|---|---|---|
| True shooting | 46.0%Better result for this metric. | 45.0% |
| Rebound rate | 5.6% | Unavailable |
| Assist rate | 16.2%Better result for this metric. | 6.4% |
| Steal rate | 2.1% | Unavailable |
| Block rate | 0.3% | Unavailable |
| Turnover rate | 14.2% | Unavailable |
Basketball context, not inputs to the regular-season impact ranking.
| Brian Oliver | Measure | Joe Kennedy |
|---|---|---|
Regular-season team contextFinal records for teams in the seasons the player represented them. These are team totals, not wins in the games he personally played. A traded season can include more than one team. Brian Oliver leads 2; Joe Kennedy leads 3; 0 tied. | ||
| 150 | Team regular-season wins | 164Leads this measure. |
| 45.7%Leads this measure. | Team winning percentage | 0.4% |
| 1 | 50-win team seasons | 2Leads this measure. |
| 0 | 60-win team seasons | 2Leads this measure. |
| 2Leads this measure. | Playoff team seasons | 0 |
Playoff contextTeam playoff results count only seasons when the player appeared in a playoff game. These are team achievements, not a measure of individual playoff impact. Brian Oliver leads 0; Joe Kennedy leads 0; 0 tied; 5 unavailable or not comparable. | ||
| 1 | Playoff appearances | Unavailable |
| 4 | Playoff games | Unavailable |
| 4 | Team playoff wins | Unavailable |
| 0 | Finals appearances | Unavailable |
| 0 | Championships | Unavailable |
Source and scope: Basketball Reference, NBA seasons 1950–51 through 2025–26, accessed 2026-10-09.
Career Shape Explorer
Follow every regular season by calendar year, NBA experience or age. This view describes the path; it does not create another verdict.
Estimated points per 100 possessions above that season's league average. Zero is average; higher is better.
Some seasons have fewer than 40 games. Select a season to see each player's games and the size of the sample.
Calendar years: both players share the same season scale.
Brian Oliver: peak unavailable
Brian Oliver: no recorded NBA seasons from 1992–93 through 1993–94, between 1991–92 and 1994–95. The gap remains empty on the calendar scale.
Brian Oliver: no recorded NBA seasons from 1995–96 through 1996–97, between 1994–95 and 1997–98. The gap remains empty on the calendar scale.
Historical defense is estimated. Select a season for its source qualification; an estimate is not a confidence interval.
Every observed season keeps its age, era-centered impact estimate, league rank when qualified, career-stage label, and exact evidence link. A missing career point stays missing.
Observed values come from the accepted season comparison record. Historical source qualifications stay attached to affected seasons. Missing means no NBA season exists at that aligned point; it is never interpolated. Missing is not the same as estimated.
Individual-season deep dive
These are individual seasons selected to illustrate each player's statistical profile. They are not career averages, peak-window averages, or the calculation behind the comparison verdict.
Starts with each player’s best qualifying published season rank, using the latest year to break ties. Change either year without changing the verdict above.
· Loading profile
· Loading profile
Reading the bars: Actual values come first. Percentiles show standing within a season; higher usually means a higher value. Blue and purple identify the players, while gray describes role or workload.
Each measure’s help button also provides its exact population. Impact Rate is a published 0–100 score, not a percentile. Turnover avoidance reverses the direction: higher means fewer turnovers. Percentiles do not measure era strength or add points to the verdict. Missing data are not zero.
Loading same-season percentile…
Loading same-season percentile…
Loading same-season percentile…
Loading same-season percentile…
Loading same-season percentile…
Loading same-season percentile…
Loading same-season percentile…
Loading same-season percentile…
Loading same-season percentile…
Loading same-season percentile…
Loading same-season percentile…
Loading same-season percentile…
Loading same-season percentile…
Loading same-season percentile…
Loading same-season percentile…
Loading same-season percentile…
Loading same-season percentile…
Loading same-season percentile…
Loading same-season percentile…
Loading same-season percentile…
Loading same-season percentile…
Loading same-season percentile…
Loading same-season percentile…
Loading same-season percentile…