hussEyquation
Menu

NBA player comparison

Bo Ellis vs. Tom Kozelko

Change players
No shared season availableCareerPeak
More ways to compare

Career comparison · Regular season

Tom Kozelko ranks higher in career impact

Bo Ellis

1977–78 to 1979–80 · 3 seasons

No. 1912 Career Impact Rank
207.8733 Career Composite Lower is better

Tom Kozelko

1973–74 to 1975–76 · 3 seasons

No. 1582 Career Impact Rank
163.8400 Career Composite Lower is better

Tom Kozelko's best-3 average finish (155 vs 198.1667) and best-5 average finish (168.6 vs 213.1) and best-10 average finish (178.8 vs 224.3) gives the lower Composite after weighting. The Career Composite gap is 44.0333; 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 Bo Ellis and Tom Kozelko. The calculation below shows the published replacement rule; it does not predict how a career will finish.

Historical defense is estimated. See source qualifications.
Bo Ellis: Recorded box inputsIncludes bounded historical defense estimate
Tom Kozelko: Recorded box inputsIncludes bounded historical defense estimate

The ranking calculation Average league finish · lower is better

PlayerBest 350% weightBest 530% weightBest 1020% weight
Bo Ellis198.17213.10224.30
Tom Kozelko155.00168.60178.80

These three windows determine Career Impact Rank. See the exact calculation →

The seasons behind their rankings

The same ranking years, viewed through their basketball statistics. Descriptive seasonal averages, not additional ranking inputs.

50% Best 3 + 30% Best 5 + 20% Best 10

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.

How to read the percentile bars

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.

See included seasons and weights

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.

Bo Ellis

  1. 1979–80No. 161 · 24.67%Best 3 · Best 5 · Best 10
  2. 1977–78No. 198 · 24.67%Best 3 · Best 5 · Best 10

Tom Kozelko

  1. 1973–74No. 143 · 24.67%Best 3 · Best 5 · Best 10
  2. 1975–76No. 148 · 24.67%Best 3 · Best 5 · Best 10
  3. 1974–75No. 174 · 24.67%Best 3 · Best 5 · Best 10

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.

Uses the selected window. Download social card and Share image use this detailed graphic too.

Select a complete common ranking window and let its source comparisons finish loading to preview a graphic.

Career totals, peak and full comparison table

The comparison in numbers

Ranking and impact

Bo EllisMeasureTom Kozelko
Career Impact RankPeak-weighted regular-season rankLower is better.
Bo Ellis#1912
Tom Kozelko#1582Better result for this measure.
Best 3 average finish50% weightLower is better.
Bo Ellis198.2
Tom Kozelko155.0Better result for this measure.
Best 5 average finish30% weightLower is better.
Bo Ellis213.1
Tom Kozelko168.6Better result for this measure.
Best 10 average finish20% weightLower is better.
Bo Ellis224.3
Tom Kozelko178.8Better result for this measure.
Peak era-centered impactEstimated points per 100 possessions above that season's league average
Bo Ellis-1.5Better result for this measure.
Tom Kozelko-2.3
Points above replacementRegular season · fixed -2.0 replacement baseline · higher is better
Bo Ellis-21 ptsBetter result for this measure.
Tom Kozelko-35 pts

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.

Career scoring, passing and playing time

Bo EllisMeasureTom Kozelko
Games
Bo Ellis168
Tom Kozelko189
Minutes
Bo Ellis1,984
Tom Kozelko1,911
Points per game
Bo Ellis3.6Better result for this measure.
Tom Kozelko2.2
Rebounds per game
Bo Ellis2.9Better result for this measure.
Tom Kozelko1.8
Assists per game
Bo Ellis0.7Better result for this measure.
Tom Kozelko0.5
True shooting
Bo Ellis48.4%Better result for this measure.
Tom Kozelko46.0%

Green marks the better result for a directional measure. Playing time describes workload, not quality. Missing data are not zero.

Share

Why they rank this way

Tom Kozelko'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.

Tom Kozelko's advantages contribute 44.0333 Composite points, offset by 0 in the other direction. The net difference is 44.0333. Displayed values are rounded; the calculation retains full precision.

Bo Ellis had the higher single-season impact estimate and more total career value.

Regular season only. Impact figures are estimates; small gaps are not exact. Playoffs, awards and titles do not determine the verdict.

The strongest case for each player

The measured case for Bo Ellis

Points-above-replacement rank: #3292, compared with #3781. Peak era-centered impact: -1.5, compared with -2.3. Points above replacement: -21 pts, compared with -35 pts. These are distinct measures, not votes or additional points added to the ranking.

The measured case for Tom Kozelko

Best 3 average finish: 155.0, compared with 198.2. Best 5 average finish: 168.6, compared with 213.1. Best 10 average finish: 178.8, compared with 224.3. These are distinct measures, not votes or additional points added to the ranking.

Career Impact RankSee exactly how both rankings are calculatedBo Ellis207.873 Composite · Lower is better
No. 1,912 all-time
Tom Kozelko163.840 Composite · Lower is better
No. 1,582 all-time

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.

Bo Ellis

207.873 Composite · Lower is better · No. 1,912 all-time

Bo Ellis's Career Composite windows
Window and finishes usedAverage finish
Lower is better
WeightContribution
Lower is better
Best 3No. 1611979–80Recorded box inputsIncludes bounded historical defense estimateNo. 1981977–78Recorded box inputsIncludes bounded historical defense estimate1 open slot × No. 235.5198.16750%99.083
Best 5No. 1611979–80Recorded box inputsIncludes bounded historical defense estimateNo. 1981977–78Recorded box inputsIncludes bounded historical defense estimate3 open slots × No. 235.5213.10030%63.930
Best 10No. 1611979–80Recorded box inputsIncludes bounded historical defense estimateNo. 1981977–78Recorded box inputsIncludes bounded historical defense estimate8 open slots × No. 235.5224.30020%44.860

99.083 + 63.930 + 44.860= 207.873

Open-season rule: every unfilled slot uses No. 235.5, this player's stored replacement finish.

Tom Kozelko

163.840 Composite · Lower is better · No. 1,582 all-time

Tom Kozelko's Career Composite windows
Window and finishes usedAverage finish
Lower is better
WeightContribution
Lower is better
Best 3No. 1431973–74Recorded box inputsIncludes bounded historical defense estimateNo. 1481975–76Recorded box inputsIncludes bounded historical defense estimateNo. 1741974–75Recorded box inputsIncludes bounded historical defense estimate155.00050%77.500
Best 5No. 1431973–74Recorded box inputsIncludes bounded historical defense estimateNo. 1481975–76Recorded box inputsIncludes bounded historical defense estimateNo. 1741974–75Recorded box inputsIncludes bounded historical defense estimate2 open slots × No. 189168.60030%50.580
Best 10No. 1431973–74Recorded box inputsIncludes bounded historical defense estimateNo. 1481975–76Recorded box inputsIncludes bounded historical defense estimateNo. 1741974–75Recorded box inputsIncludes bounded historical defense estimate7 open slots × No. 189178.80020%35.760

77.500 + 50.580 + 35.760= 163.840

Open-season rule: every unfilled slot uses No. 189, 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 →

Supporting evidence

Individual statistics across the best ten impact seasons

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.

Bo Ellis: 2 qualifying seasons. Tom Kozelko: 3 qualifying seasons.

Best-ten-season statistical context: Bo Ellis and Tom Kozelko
MetricBo EllisTom Kozelko
True shooting48.5%Better result for this metric.46.8%
Rebound rate12.4%Better result for this metric.9.0%
Assist rate7.8%Better result for this metric.6.5%
Steal rate1.4%1.6%Better result for this metric.
Block rate2.3%Better result for this metric.0.5%
Turnover rate17.2%Unavailable

Winning, playoffs, and awards

Basketball context, not inputs to the regular-season impact ranking.

Career team results and honors
Bo EllisMeasureTom Kozelko

Regular-season team context

Final 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.

Bo Ellis leads 1; Tom Kozelko leads 4; 0 tied.

125Team regular-season wins246Leads this measure.
50.8%Leads this measure.Team winning percentage0.4%
050-win team seasons3Leads this measure.
060-win team seasons3Leads this measure.
2Playoff team seasons3Leads this measure.

Playoff context

Team playoff results count only seasons when the player appeared in a playoff game. These are team achievements, not a measure of individual playoff impact.

Bo Ellis leads 0; Tom Kozelko leads 4; 1 tied.

2Playoff appearances3Leads this measure.
15Playoff games25Leads this measure.
7Team playoff wins14Leads this measure.
0Finals appearances1Leads this measure.
0Championships0
How these context measures are counted
Team regular-season wins and winning percentage
Wins credited to each listed team's final record in every season the player represented that team. Winning percentage divides those wins by those teams' total games. These do not measure wins personally created by the player.
50- and 60-win team seasons
Listed team-seasons finishing with at least 50 or 60 wins. Shortened seasons are not projected to 82 games.
Team seasons reaching playoffs
Listed team-seasons in which the team qualified for the playoffs.
Playoff appearances and games
Postseasons and games in which the player appeared.
Team playoff wins, Finals and championships
Team results in postseasons when the player appeared. Finals and championships count seasons when he played in the playoffs for the finalist or champion.
Awards
Recorded NBA selections and major regular-season honors. Tied counts remain ties. Missing source fields are not zero.

Source and scope: Basketball Reference, NBA seasons 1950–51 through 2025–26, accessed 2026-10-09.

Career Shape Explorer

How the careers developed

Follow every regular season by calendar year, NBA experience or age. This view describes the path; it does not create another verdict.

View
Align careers by
Bo EllisTom Kozelko◆ Single-season peak

Estimated points per 100 possessions above that season's league average. Zero is average; higher is better.

Calendar years: both players share the same season scale.

Tom Kozelko: peak unavailable

Historical defense is estimated. Select a season for its source qualification; an estimate is not a confidence interval.

Axis years are season-ending years: ’82 means 1981–82. The full observed range stays visible. Select a bar, use the season selector, or use keyboard arrow keys for exact values. Empty positions are not zero. Season timeline contains the complete linked evidence.
Explore individual seasons: full statistical profiles

Individual-season deep dive

Explore individual seasons

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.

Bo Ellis

· Loading profile

Tom Kozelko

· 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.

Exact comparison groups and eligibility

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.

Impact Rate
Published model score, not a percentile.
Season Value
Comparison population unavailable.
Offensive lineup impact
Comparison population unavailable.
Defensive lineup impact
Comparison population unavailable.
True shooting
Comparison population unavailable.
Field-goal shooting
Comparison population unavailable.
Three-point shooting
Comparison population unavailable.
Free-throw shooting
Comparison population unavailable.
Free-throw rate
Comparison population unavailable.
Three-point attempt rate
Comparison population unavailable.
Assist rate
Comparison population unavailable.
Turnover avoidance
Comparison population unavailable.
Rebound rate
Comparison population unavailable.
Steal rate
Comparison population unavailable.
Block rate
Comparison population unavailable.
Offensive usage
Comparison population unavailable.
Points per game
Comparison population unavailable.
Rebounds per game
Comparison population unavailable.
Assists per game
Comparison population unavailable.
Minutes per game
Comparison population unavailable.
Games played
Comparison population unavailable.
Total minutes
Comparison population unavailable.

Impact

Impact RatePublished 0–100 score · higher is better
Bo EllisNo seasonLoading…

Loading same-season percentile…

Tom KozelkoNo seasonLoading…

Loading same-season percentile…

Season Value
Bo EllisNo seasonLoading…

Loading same-season percentile…

Tom KozelkoNo seasonLoading…

Loading same-season percentile…

Offensive lineup impact
Bo EllisNo seasonLoading…

Loading same-season percentile…

Tom KozelkoNo seasonLoading…

Loading same-season percentile…

Defensive lineup impact
Bo EllisNo seasonLoading…

Loading same-season percentile…

Tom KozelkoNo seasonLoading…

Loading same-season percentile…

Scoring

True shooting
Bo EllisNo seasonLoading…

Loading same-season percentile…

Tom KozelkoNo seasonLoading…

Loading same-season percentile…

Passing and possession

Assist rate
Bo EllisNo seasonLoading…

Loading same-season percentile…

Tom KozelkoNo seasonLoading…

Loading same-season percentile…

Turnover avoidanceHigher percentile = fewer turnovers
Bo EllisNo seasonLoading…

Loading same-season percentile…

Tom KozelkoNo seasonLoading…

Loading same-season percentile…

Rebound rate
Bo EllisNo seasonLoading…

Loading same-season percentile…

Tom KozelkoNo seasonLoading…

Loading same-season percentile…

Role and playing time

Offensive usagePossessions finished with shots, free throws or turnovers. Not touches or passing involvement.Role / volume · not a quality score
Bo EllisNo seasonLoading…

Loading same-season percentile…

Tom KozelkoNo seasonLoading…

Loading same-season percentile…

Points per gameRole / volume · not a quality score
Bo EllisNo seasonLoading…

Loading same-season percentile…

Tom KozelkoNo seasonLoading…

Loading same-season percentile…

Games playedRole / volume · not a quality score
Bo EllisNo seasonLoading…

Loading same-season percentile…

Tom KozelkoNo seasonLoading…

Loading same-season percentile…

Total minutesRole / volume · not a quality score
Bo EllisNo seasonLoading…

Loading same-season percentile…

Tom KozelkoNo seasonLoading…

Loading same-season percentile…

Current-season data updates