The following is an analysis of the video game Super Mario Strikers for Gamecube. Data was obtained by manual input at the conclusion of each individual game on an excel spreadsheet. Data was collected for 2,400 games. 426 games were recorded against the computer while the remaining 1,974 games were human against human games. Each game was set to last two minutes each, for the exception of overtime.
As shown below, some stats have been adjusted in order to represent a two minute average so as to eliminate any bias from games that reached overtime. In overtime, games would continue past the two minute mark until a go-ahead goal was scored to determine a winner. For example, a two minute game could last as long as four minutes since a goal was not scored for two minutes past regulation time. In many instances, this created inflated numbers for shots taken and hits while deflating averages for goals scored.
This is the full data set. Here you are able to sort through the data yourself to see exactly what data is being used to create the following analysis.
In order to provide a general baseline of performance in a typical game, this table shows typical expectations from each game.
| Goals_2min | Shots_2min | Shot_Efficiency | |
|---|---|---|---|
| Min. :0.0000 | Min. : 0.7947 | Min. :0.00000 | |
| 1st Qu.:0.8571 | 1st Qu.: 6.0000 | 1st Qu.:0.09091 | |
| Median :1.0000 | Median : 8.0000 | Median :0.18182 | |
| Mean :1.5475 | Mean : 7.8130 | Mean :0.19511 | |
| 3rd Qu.:2.0000 | 3rd Qu.: 9.8212 | 3rd Qu.:0.28571 | |
| Max. :8.0000 | Max. :18.0000 | Max. :1.00000 |
| Stat | Formula | Explanation |
|---|---|---|
| Shot Efficiency | Goals/Shots | Average number of goals made per shot taken. |
| Goals_2min | (Goals*120)/Time | Average number of goals per 2 minute game. |
| Hits_2min | (Hits*120)/Time | Average number of hits per 2 min. |
| Steals_2min | (Steals*120)/Time | Average number of steals per 2 min. |
| PerfP_2min | (Perfect P*120)/Time | Average number of Perfect Passes per 2 min. |
| GoalsA_2min | (GoalsA*120)/Time | Average goals allowed per 2 min. |
| Goal_Diff_2min | Goals_2min-GoalsA_2min | Average goal differential per 2 min. |
| Shots_2min | (Shots*120)/Time | Average number of shots per 2 min. |
| Shot_Efficiency_Against | GoalsA/ShotsA | Shot efficiency by opposing team(s). |
Below is a table of individual player stats against one another.
| Player | Opponent | Games | Wins | Losses | Win_Perc | GoalsPG | Shot_Eff | GAgainst | Goal_Differential |
|---|---|---|---|---|---|---|---|---|---|
| Preston | Josh | 16 | 15 | 1 | 0.938 | 2.640 | 0.230 | 0.729 | 1.911 |
| Preston | Matt | 631 | 431 | 200 | 0.683 | 2.067 | 0.227 | 1.270 | 0.797 |
| Preston | Brandon | 22 | 14 | 8 | 0.636 | 1.614 | 0.186 | 1.059 | 0.555 |
| Brandon | Matt | 136 | 86 | 50 | 0.632 | 1.510 | 0.182 | 1.108 | 0.402 |
| Matt | Josh | 182 | 114 | 68 | 0.626 | 1.603 | 0.179 | 1.169 | 0.434 |
| Josh | Matt | 182 | 68 | 114 | 0.374 | 1.169 | 0.175 | 1.603 | -0.434 |
| Matt | Brandon | 136 | 50 | 86 | 0.368 | 1.108 | 0.148 | 1.510 | -0.402 |
| Brandon | Preston | 22 | 8 | 14 | 0.364 | 1.059 | 0.142 | 1.614 | -0.555 |
| Matt | Preston | 631 | 200 | 431 | 0.317 | 1.270 | 0.173 | 2.067 | -0.797 |
| Josh | Preston | 16 | 1 | 15 | 0.062 | 0.729 | 0.120 | 2.640 | -1.911 |
This table shows, simply based on Win Percentage, that Preston is the most dominant player. Preston claims the top 3 spots for Win Percentage against each of his opponents. In fact, he has some of the best stats of any player with data available.
This next table shows individual stats regardless of opponent, which also shows Preston’s dominance.
| Player | Games | Wins | Losses | Win_Perc | GoalsPG | Shot_Eff | GAgainst | Goal_Differential |
|---|---|---|---|---|---|---|---|---|
| Preston | 669 | 460 | 209 | 0.688 | 2.065 | 0.226 | 1.250 | 0.815 |
| Brandon | 158 | 94 | 64 | 0.595 | 1.447 | 0.177 | 1.178 | 0.269 |
| Matt | 949 | 364 | 585 | 0.384 | 1.310 | 0.171 | 1.815 | -0.504 |
| Josh | 198 | 69 | 129 | 0.348 | 1.133 | 0.171 | 1.687 | -0.553 |
Overall, it is pretty clear, simply from numbers, that Preston is the better player.
CPU’s provide a fairly consistent gameplay that help to show how strong certain players are. In this case, we only have two players that have played against the CPU to provide accurate data. But the evidence still shows how much better Preston is than any individual.
| Player | Opponent | Games | Wins | Losses | Win_Perc | GoalsPG | ShotsPG | Shot_Eff | GAgainst | Goal_Differential |
|---|---|---|---|---|---|---|---|---|---|---|
| CPU | Preston | 78 | 3 | 75 | 0.038 | 0.540 | 3.631 | 0.152 | 2.983 | -2.443 |
| CPU | Matt | 135 | 42 | 93 | 0.311 | 0.941 | 5.558 | 0.169 | 1.750 | -0.809 |
During gameplay, it strongly appears that some goals are different sizes than others. This analysis was designed to show if there really was any statistically significant difference in each stadium.
Disappointingly, there was no significant difference by stadium when comparing goals and shots per two minutes. Although, there is some variance to be aware of when comparing their goal averages.
| Stadium | Games | GoalsPerGame | Shot_Efficiency | ShotsPerGame |
|---|---|---|---|---|
| Crater | 202 | 1.961 | 0.227 | 8.556 |
| Pipeline | 212 | 1.700 | 0.222 | 7.632 |
| Underground | 344 | 1.569 | 0.188 | 8.187 |
| Konga | 418 | 1.534 | 0.190 | 7.926 |
| Bowser | 322 | 1.497 | 0.183 | 8.092 |
| Battle Dome | 158 | 1.486 | 0.182 | 8.041 |
| Palace | 318 | 1.333 | 0.171 | 7.716 |
Using this table, it does appear there is some variance between the stadiums. Crater easily has the best shot efficiency and goals per game while Palace has the lowest in both categories. It may be interesting to test this theory further with longer length games to gain a better picture of the difference in the stadiums.
Players that win have different ways of playing to those that lose. Let’s look at how shot per two minutes and goals relate for games where a player wins vs when a player loses. The first visual shows losses while the second shows a graphic for wins based.
While wins show some consistency, losses, particularly between 0 and 1 goal per minute, show significant differences in how goals are scored. There appears to be little relationship between shots and goals in this interval.
Based on the research conducted thus far, there is little to conclude on the influence of goals. We have been able to determine that stadiums have no statistically significant impact on goal scoring. The stat most connected with goal scoring throughout research was shown to be shot efficiency, but it has yet to be determined what makes an individual more efficient with their shots.
This bar graph shows the frequency of goals scored in a game.
This bar graph shows the frequency of goals based on wins.
This bar graph shows the frequency of goals based on losses.
The following shows a box plot of goals scored by players. Unsurprisingly, Preston’s box extends much further than the other players due to his goal scoring prowess.
This boxplot shows the data of goals scored by stadium. Interesting, the box plot for Crater field extends much further than any of the other boxes shown indicating there may be some impact on number of goals scored in that stadium.