| Course | MT241 Sport Analytics |
|---|---|
| Unit | Unit 3 |
| Paper type | Descriptive analysis |
| Length | About 1,170 words, 7 pages double-spaced |
| Format | APA 7 student paper |
| School | Purdue University Global |
| Program | BS in Business Administration |
| Updated | October 2026 |
Free sample paper for MT241 Unit 3
Ten Wins in Twelve and a Save Percentage of .939: A Unit 3 Descriptive Analysis of One Club's Season
[Student Name]
Purdue University Global
MT241: Sport Analytics
Unit 3 Descriptive Analysis
[Instructor Name]
[Date]
The club, its 82-game log and every figure are composite and illustrative; the script is a sample written for this paper; the sources cited are real.
Purpose and Data
This analysis describes our composite club's season at 5-on-5 before any claim is made about why it went the way it did. The data are the club's 82-game log, one row per game, with shot attempts for and against, expected goals for and against from one public model, goals for and against, and shots on goal for and against, all at 5-on-5. Every number in this paper was produced by the script in the appendix, so a reader can rerun it on the same file and get the same figures. Tukey (1977) described exploratory work as looking at data before asking it to confirm anything, and that is the order followed here.
The Season in Summary
Across 82 games the club took 49.2 percent of 5-on-5 shot attempts and 50.8 percent of 5-on-5 expected goals. It was, in other words, slightly outshot but created chances of slightly better quality than it allowed. The median game's expected-goals share was 50.9 percent, close to the season figure, which tells us the season was not pulled up or down by a handful of extreme games.
The spread between games was wide. The standard deviation of expected-goals share across games was 9.8 points, meaning a typical game landed anywhere from about 41 to 61 percent. A single game, or even a week of games, therefore says little about the club's level. That spread is the most useful number in this section, because it sets how long a stretch has to be before it means something.
The Game-by-Game Chart
The chart plots each game's expected-goals share as a dot, with a line showing the ten-game rolling mean. The rolling mean stays between about 47 and 54 percent for nearly the whole season. It dips in late November, sits near 48 percent through most of December, and climbs back above 51 percent in January. The twelve-game December run is shaded. Seen this way, the run is not a high point. It is one of the club's weaker stretches by chances.
The December Run Against the Rest
The segment table sets the twelve-game run beside the other seventy games.
Expected-goals share: 48.4 percent in the run, 51.2 percent in the other seventy games
Shooting percentage at 5-on-5: 10.3 percent in the run, 7.8 percent otherwise
Save percentage at 5-on-5: .939 in the run, .916 otherwise
PDO, the sum of the two: 104.2 in the run, 99.4 otherwise
During the run, the club's share of chances was nearly three points lower than its usual level. What changed was what happened to the shots once they were taken. The club's shooters scored on 10.3 percent of their shots instead of 7.8, and its goaltenders stopped 93.9 percent instead of 91.6.
PDO adds the two percentages. Across a league, it averages about 100, because every goal scored is a goal allowed by someone, and teams that run well above or below 100 for a stretch usually drift back toward it. Alamar (2013) notes that separating the parts of performance a team controls from the parts that vary by chance is one of the main jobs of analytics, and PDO is a quick way to do that in hockey.
The December run was built on finishing and saves, not on a better share of the play.
What the Columnist Saw
The columnist was right that the club won ten of twelve games. That is a real result and it earned real standings points. The claim that the team had found its identity, though, implies a lasting change in how it plays, and the chances do not show one. If the club had changed its play, the expected-goals share would have risen during the run. It fell instead. Tango et al. (2007), writing about baseball, show how often short hot stretches reflect the ordinary spread of results around a team's true level, and the same reasoning applies to twelve hockey games.
After the Run
January supports the same reading. Once the run ended, the club's shooting percentage fell back to about 7.6 percent and its save percentage to about .915, and the club lost six of its next ten games. Its expected-goals share in those ten games was 51.9 percent, higher than during the run. The team was playing slightly better by chances and winning less, which is what a return of PDO toward 100 would predict. A reader who saw only the standings would think the team had collapsed in January; the chances say it had improved.
What This Analysis Does Not Show
A descriptive analysis summarizes. It does not explain. This paper does not show why the club's shooting and goaltending were high in December: it could be partly skill, such as a goaltender playing his best stretch, and partly chance. It also does not show the run was worthless. Points won in December count the same as points won in April. What it shows is that the club's underlying share of chances was no better during the run than at other times, so planning for the rest of the season should start from the 51 percent club the full season describes, not from the 83 percent winning rate of December.
Limits
Expected goals here come from one public model; another model would give slightly different numbers, though the direction of the comparison is unlikely to change. The analysis uses 5-on-5 play only and leaves out power plays and penalty kills, which also affected December results. Twelve games is a small sample for any of these rates, which is one more reason not to read too much into them.
Appendix: The Script
The script is about forty lines of Python using the pandas library (McKinney, 2022). It runs in five steps.
1. It reads the game log into a table and checks that it has 82 rows and no missing values in the shot, expected-goals and goal columns.
2. It computes each game's shares, such as expected goals for divided by expected goals for plus against, and the season totals, summing the columns before dividing so that each game is weighted by its volume.
3. It reports the season shares, the median game and the standard deviation of game shares.
4. It flags the twelve December games by date, then computes shooting percentage, save percentage, PDO and expected-goals share separately for the run and for the other seventy games.
5. It draws the game-by-game chart with a ten-game rolling mean and shades the run.
The script's output is pasted into the tables above unchanged. Anyone rerunning it on the same file should get the same figures to one decimal place.
References
Alamar, B. C. (2013). Sports analytics: A guide for coaches, managers, and other decision makers. Columbia University Press.
McKinney, W. (2022). Python for data analysis: Data wrangling with pandas, NumPy, and Jupyter (3rd ed.). O'Reilly Media.
Tango, T. M., Lichtman, M. G., & Dolphin, A. E. (2007). The book: Playing the percentages in baseball. Potomac Books.
Tukey, J. W. (1977). Exploratory data analysis. Addison-Wesley.
The Purdue Global MT241 Unit 3 instructions
The third unit of MT 241 asks students to describe a data set before drawing conclusions from it. Summarize the whole set first with measures of center and spread, then compare a segment of interest, such as a streak, a coaching change or a stretch of home games, against the rest. The prompt expects every number to be reproducible, so many sections require the code or spreadsheet steps that produced the figures, either in an appendix or as an attachment. Include at least one chart and one comparison table, and say clearly what the description does not show. Five pages is common, cited in APA 7 with the course text and a source on the method, and composite data should be labeled.
Inside the MT241 Unit 3 example
A purpose and data section names the 82-game log and the one public expected-goals model, with Tukey's order of looking before confirming. The season summary gives the shares, the median game and the spread, and a margin note marks the spread as the yardstick for the streak. A chart section describes the ten-game rolling mean and the shaded run. The four-row segment table sets shooting, saves and PDO for the run and the rest, and the highlighted line places the run on finishing and saves. The columnist's claim is tested against Tango and colleagues' work on hot stretches, and an after-the-run section shows January's better chances and worse results. Limits, then an appendix describing the script in five steps, close it with a final note.
MT241 Unit 3 rubric: what earns full points
Graders of a descriptive analysis tend to look for accuracy, reproducibility, sound comparison and restraint. Correct calculations that match the tables and charts earn the base marks. Reproducibility earns more when the script or steps are included and a reader could rerun them. The comparison scores well when the segment is set against a baseline from the same data, with spread reported so the reader can judge size. Restraint matters: a paper that claims causes from a description loses points, while one that states what it cannot show gains them. Clear charts with labeled axes, a short limits section and correct citations for the method sources complete the rubric.
Common MT241 Unit 3 mistakes, and how to avoid them
Students often lead with the interesting streak and never describe the full season, which leaves no baseline. Another common gap is reporting averages without spread, so a reader cannot tell whether a difference is large. Some papers include code that does not match the tables, or none at all when the prompt asks for it. Others slip into explanation, saying a team found its identity, when the unit asks only for description. Weight season totals by volume rather than averaging game percentages. If you have a data set but no script yet, a writer can draft the steps with you so every figure in the tables can be rerun from the file.
Get an MT241 Unit 3 example written to your instructions
Send the season, team or dataset your Unit 3 prompt names, whether a game log it supplies or one you choose, and the rubric. The first custom analysis costs nothing and arrives in 24-48h, the whole season described before any run is singled out, and a short script included where the course expects figures computed rather than typed. The paper above is an original model document written by our desk, not a submitted student paper and not an official Purdue University Global document.
More MT241 papers and BS in Business Administration samples
- MT241 Unit 1: Discussion Board Post
- MT241 Unit 2: Metric Definition Exercise
- MT241 Unit 4: Player Comparison Report
- MT241 Unit 5: Data Quality Review
- MT241 Unit 6: Seminar Reflection
- MT241 Unit 7: Visualization Exercise
- MT241 Unit 8: Tactical Analysis
- MT241 Unit 9: Recruitment Recommendation
- MT241 Unit 10: Analytics Report
- MT240 Unit 3: Media Representation Analysis
- MT237 Unit 3: Information Flow Analysis
- MT236 Unit 3: Supplier Evaluation
- MT221 Unit 3: Customer Expectations Report
MT241 Unit 3 questions, answered
What is PDO in hockey analytics?
The sum of a team's or player's shooting percentage and save percentage, usually at 5-on-5, on a scale where 100 sits near league average. Values far above or below 100 over short stretches tend to move back toward it, because neither percentage is fully controlled by the skaters. Analysts use it as a flag for luck, not as a measure of quality.
Why include the script in a descriptive analysis?
Because MT241 treats analysis as something another person should be able to reproduce. A short script, in Python, R or a spreadsheet with visible formulas, shows how each number was made and lets a grader check a filter or a denominator. The example prints forty lines in an appendix; some sections ask for the file itself as an attachment.
How long does a stretch have to be before it means something?
Longer than most fans expect. In hockey, shot-attempt share settles faster than goal-based measures because it counts many more events, while shooting and save percentages need far larger samples. A twelve-game run can show a real change in process if shot or chance share moves, but results alone over that length mostly reflect variation.
Where can I find a free MT241 Unit 3 sample paper?
You can read the complete MT241 Unit 3 descriptive analysis on this page. It summarizes an 82-game season, compares a twelve-game December run with the rest and describes the script that produces each figure.
What does PDO show in the MT241 Unit 3 sample?
PDO adds shooting and save percentage. It reads 104.2 during the December run against 99.4 for the rest of the season, which tells the reader the streak rested on finishing and goaltending rather than a better share of chances.