A basketball table can contain accurate percentages and still invite a wrong comparison. One player makes one of two shots, another makes nine of eighteen, and both show 50%. In that particular case, averaging their percentages works. Change the second line to nine of ten, however, and the same shortcut fails. The average of 50% and 90% is 70%, while the combined shooting is ten makes from twelve attempts: about 83.3%.
The difference is not a statistical trick. The two players supplied different numbers of attempts. Before combining percentages, return to the counts that produced them. This guide uses invented basketball examples to show how to combine rows, compare like with like and describe the result without turning a small sample into a verdict on a player.
Read what the columns actually contain
Field-goal tables usually pair makes and attempts, sometimes in separate columns and sometimes as a line such as 7–16. Check the heading before interpreting the dash. A score, a date and a makes–attempts line can all contain two numbers; their meanings come from the labels.
The NBA glossary defines field-goal percentage as makes divided by attempts. To express that ratio as a percentage, multiply by 100. Thus seven divided by sixteen is 0.4375, or 43.75%. A display rounded to one decimal place would show 43.8%.
The FIBA Statisticians' Manual 2024, version 2.0 distinguishes credited field-goal attempts from made field goals and explains that a make also counts as an attempt. It includes situation-specific recording rules, so a viewer should not replace official totals merely by counting every visible shooting motion. This article reads already-recorded totals; it is not training for deciding how an official statistician should credit a disputed play.
Use a denominator before using a ranking
Consider two fictional shooting lines from one game. Ari makes four of five attempts. Bo makes seven of fifteen. Ari's percentage is 80%; Bo's is approximately 46.7%. Ari has the higher conversion percentage in this sample. Bo has more made field goals. Those statements answer different questions and can both be true.
A comparison becomes misleading when its wording changes halfway through. “Ari had the higher percentage” is supported by the counts. “Ari contributed more to the team” requires a much broader evaluation. Even “Ari scored more” is not established by these field-goal counts alone, because the types of made shots and free throws have not been supplied.
Write the count beside the percentage whenever practical: “4 of 5, 80%” carries more information than “80%” alone. The denominator tells the reader how much recorded opportunity produced the figure. It also makes the arithmetic checkable without requiring access to your spreadsheet.
Combine makes and attempts separately
Now combine Ari and Bo's lines. Four plus seven gives eleven makes. Five plus fifteen gives twenty attempts. The combined percentage is therefore eleven divided by twenty, or 55%. Averaging 80% and 46.7% gives about 63.35%, which is not the percentage of their combined attempts.
Why does the shortcut fail? It gives Ari's five attempts the same total influence as Bo's fifteen. In the combined attempt pool, Ari supplied one quarter of the attempts and Bo supplied three quarters. The higher-volume row must therefore have more influence on the combined conversion rate.
You can reproduce the correct result with weights: one quarter of 80% plus three quarters of the unrounded seven-fifteenths percentage equals 55%. But when the raw counts are available, adding them is simpler and avoids rounding complications. Keep a makes total and an attempts total, then calculate the percentage once.
The same method applies across games, periods or players, provided the rows describe compatible and non-overlapping observations. A correct formula cannot rescue a dataset that mixes unrelated competitions or counts the same game twice.
Do not average the displayed percentages
Suppose a fictional player goes two for three in one game and three for seven in the next. The combined record is five for ten, exactly 50%. The two individual percentages are approximately 66.7% and 42.9%. Their simple average is approximately 54.8%, not the combined percentage.
There are two separate issues here. First, each game contains a different number of attempts. Second, the displayed percentages are rounded. Weighting the rounded values by three and seven produces a value close to, but not exactly, 50%. Using the integer counts avoids both problems.
This matters when copying statistics into a season summary. A column of attractive percentages is not a replacement for the underlying makes and attempts. If a source gives only percentages without denominators, you may be unable to reconstruct the combined figure reliably. Leave that result unavailable rather than inventing equal weights.
An average of game percentages can be calculated deliberately for a different analytical question. If you do that, label it precisely. Do not call it the player's overall shooting percentage, because it treats a low-attempt game and a high-attempt game as equally weighted observations.
Keep overlapping categories out of the total
A common reading error is to treat the three-point column as extra shooting outside the overall field-goal line. In an overall field-goal summary, the made threes are already included among the made field goals. Before adding columns, establish whether they are separate categories or whether one is contained inside another.
Here is an invented complete line: seven made field goals from sixteen attempts, including three made threes from eight attempts, plus four made free throws from five attempts. Subtracting the three-point line from the overall field-goal line leaves four made two-point shots from eight attempts.
The points can be checked as four times two, plus three times three, plus four: 8 + 9 + 4 = 21. Alternatively, count every made field goal as two points first, then add one extra point for each made three and finally the free throws: 14 + 3 + 4 = 21.
Adding seven field goals and three threes as if they were ten different made baskets would double-count the threes. Similarly, adding the sixteen total field-goal attempts and eight three-point attempts would inflate the attempt count. The underlying issue is category overlap, not difficult arithmetic.
Give free throws their own denominator
In the invented line above, free-throw shooting is four from five, or 80%. Field-goal shooting is seven from sixteen, or 43.75%. Adding the makes and attempts gives eleven from twenty-one, about 52.4%, but that result should not be labelled field-goal percentage. You have created a different mixture of attempts.
The FIBA manual's free-throw section treats free-throw makes and attempts separately and describes how violations can affect the credited record. For this reader's exercise, use the published free-throw counts as given rather than guessing from a short video clip.
If your question concerns scoring rather than conversion frequency, say so. Counting made attempts and counting points are different operations. An invented player making two threes has six points from those baskets; another making three twos also has six. Their number of successful field-goal attempts differs even though this portion of their scoring is equal.
Understand a zero-attempt row
A player who takes no field-goal attempts has no field-goal conversion sample for that period. Dividing zero by zero does not produce a meaningful shooting percentage. A website may display a dash, a blank or a numeric placeholder, so inspect the makes and attempts before interpreting its presentation.
Suppose one teammate makes three of six and another takes no shots. Together they have three makes from six attempts: 50%. Treating the second player's displayed placeholder as a true 0% and averaging it with 50% would incorrectly produce 25%.
For a worksheet, keep the raw row as zero makes and zero attempts. Show the individual percentage as unavailable when attempts are zero, while allowing those zero counts to contribute normally to a combined total. The combined calculation works as long as the total denominator is positive.
A player taking no shots may have contributed in other ways. This article does not infer passivity, poor performance or a tactical role from that single absence. A missing shooting sample is a limit on the shooting comparison, not a complete description of someone's time on court.
Watch how one attempt changes a small sample
Start with an invented four-for-five line: 80%. If the next attempt misses, the line becomes four for six, approximately 66.7%. If it goes in instead, the line becomes five for six, approximately 83.3%. One event creates a large visible difference because the original denominator is small.
Now start with forty makes from fifty attempts, also 80%. One additional miss produces forty from fifty-one, approximately 78.4%. The change is much smaller. These examples demonstrate arithmetic sensitivity; they do not establish a confidence interval or estimate either player's underlying ability.
When writing a recap, a count-based sentence can prevent overstatement. “Ari made four of five attempts in this game” reports the performance directly. Calling Ari an “80% shooter” without the sample context can sound like a claim about a stable long-term rate.
More observations can make a descriptive rate less sensitive to one extra outcome, but they do not automatically make every comparison fair. The kinds of attempts, opponents and situations may still differ. State the period and scope before interpreting the number.
A combined ranking can hide different shot mixes
Imagine two fictional players whose attempts are divided into two labelled groups. In Group One, Ari makes eight of ten while Bo makes seventy-five of one hundred. Ari's rate is higher: 80% against 75%. In Group Two, Ari makes thirty of one hundred while Bo makes two of ten. Ari again has the higher rate: 30% against 20%.
Across both groups, however, Ari has thirty-eight makes from one hundred and ten attempts, about 34.5%. Bo has seventy-seven from one hundred and ten, exactly 70%. The overall order reverses because Ari took many more attempts in the lower-conversion group, while Bo took many more in the higher-conversion group.
These are deliberately constructed numbers, not a real player's shot profile. They show that aggregation can change the question being answered. The overall percentage describes each player's actual mixture in the example. The group percentages describe performance within each defined group.
Do not conclude that one of those calculations must be wrong merely because the rankings differ. Instead, inspect how the attempts are distributed. Also avoid treating the group comparison as proof of causal superiority: the invented groups omit defensive pressure, role and other conditions that would matter in a real evaluation.
Build a worksheet that exposes mistakes
Use one row per non-overlapping observation. Include the player or team, date, competition, period covered, makes and attempts. If recording two-point and three-point detail, give those their own columns and make the relationship to the overall field-goal total explicit.
A useful arithmetic check is that makes cannot exceed attempts in a conventional credited shooting line. Another is that two-point and three-point makes add to total field-goal makes, with the same relationship for attempts. A failed check means you should investigate the copied labels or source; it does not tell you which value to silently change.
Keep a source link and retrieval date next to the row. If an official record is corrected, update the affected counts and recalculate the summary. Do not merely edit the displayed percentage, leaving an inconsistent makes–attempts pair underneath it.
Before combining rows, check for duplicate games and mixed scopes. A season-to-date total already includes individual games from that season. Adding those games again duplicates them. Likewise, a full-game line already contains its quarters. A neat total can still be wrong if the input rows overlap.
Describe the answer you actually calculated
A short, useful comparison contains the scope, the count and the interpretation. For example: “Across the two invented games, the player made five of ten attempts, 50%; averaging the two displayed game percentages would answer a different question.” The reader can see both the result and the reason for the method.
For match interpretation, calculations are only one layer. The separate Trier–Vechta analysis discusses why reported shooting and rebound totals do not reconstruct every possession. Here, the narrower task is making a summary that correctly represents its own input rows.
Start with the column labels, retain the denominators and add compatible counts before dividing. Keep shot categories distinct, treat zero-attempt rows honestly and inspect the mixture behind a combined ranking. Those habits make a basketball percentage more useful without asking it to explain the whole game.
Sources and scope
Sources consulted 18 November 2026: FIBA Statisticians' Manual 2024, version 2.0, particularly field goals and free throws; NBA statistical glossary for the basic field-goal percentage definition. All player names, shooting lines and grouped comparisons in this guide are original hypothetical teaching examples. No live player ranking, match reconstruction, betting forecast or scouting assessment is claimed. The original illustration is conceptual rather than measured match data.
