A Whole Bunch of Ball Ladder Season 4 Statistics Part 2

[soccernamlak]

Based on this post which gave me some more data to play with....


In short, since player / match id is not associated yet, I did some correlation between map frequency in Season 4, game time, and average scores. So onwards...

General Facts

- 6129 Matches Played

- Average time was 9:07.

- Average score was 6-3.13

- Shortest Game? 2:24 on Cave, 6-0.

- Longest Game? 30:42 on Mayhem2, 6-5.

- Most lopsided longest game? On Core, 24:24 with final score 6-1.

- Shortest, closest game? On Cross, 3:55 with final score 6-4.

- 52.5% of the time, the right team won.

Now for fun stuff...

Stats!

First let's look at matches played.



Sadly we'll have to ignore reef and skylands for the most part as they were introduced later into the season. That being said, it is obvious that cave and planepark are the most played maps. However, the remainder of the maps hover between 334-400 games played. Considering that this range is only ~10.7% of the total games played, I'd have to say that is pretty well balanced. I'll admit it doesn't feel that way sometimes with the start_random command, but even if it doesn't do it short term, long term it works out to being an even distribution of map play time.

Next, let's look at play time:



Here, we can see minimum, maximum, and average time per map. It should be no surprise that minimum time is roughly the same on all maps: there is only so fast of a time to score 6 goals. The interesting stats are averages and maximum time to win a game.

Let's look at maximum first. It seems that Darkwar and Cross are the quickiest to finish maps overall, while Snow, Mayhem2, and Cave take a bit longer in the maximum.

Average wise, again we see darkwar coming in at bottom with maze close behind. Asteroids, Core, Funnel, and Mayhem2 take the top 4 longest average game play time.

Scores

Let's look at score frequency:



So ideally each slice should be 16.67% if evenly distributed. With that being said, 6-0 and 6-1 games are rarer than normal distribution, 6-2 is about normal, while 6-3 and up are above normal distribution, with 6-4 being the most favored outcome of a ball match. I think that hints at the effectiveness of ladder's matching system over an extended period of time, but that's only a personal observation.

[soccernamlak]

Next we'll consider if there is a trend between time played and losing side's score, with ball_cave as our focus:



Basically, there is no correlation here. Longer games don't dictate closer scoring games, and shorter games don't either.

Finally, we'll consider the distribution per map of bad (6-0/6-1), okay (6-2/6-3), and good (6-4/6-5) games:

Raw:



Percent:




Take-away here? There really isn't that many "bad" games statistically on ladder, with the most being labyrinth at 20%.

Here, also see a pretty even distribution of "good" games too across all maps, with cross being the best in that regard.



That being said, we can conclude that if length of play and closeness of game are key factors for you, then Asteroids is the map to play, with the longest average playing time and highest rate of good games. But overall, there is not one map that seems to be severely flawed in play time or game score. So despite personal preferences in maps, overall maps provide roughly equal opportunities for length of game and quality of game.

[Aki1024]

Namlak, your my hero. >D *geeks out on graphs*
Btw, I'm using these as mockups for live data. Just as soon as I have infinite coding hours (or get off Vindictus)

[Clapon]

^5

great work as always dood, would read again.

[Ribilla]

I don't particularly like the map, but this is a good argument for skylands being a suitable map for ladder as it generates the best games.

As ever, really appreciate the analysis.

[soccernamlak]

Originally Posted by Aki1024

Namlak, your my hero. >D *geeks out on graphs*
Btw, I'm using these as mockups for live data. Just as soon as I have infinite coding hours (or get off Vindictus)

I would love you forever if you got live data of all these stats.

Originally Posted by Ribilla

I don't particularly like the map, but this is a good argument for skylands being a suitable map for ladder as it generates the best games.

As ever, really appreciate the analysis.

Yep! And yeah, I was really wanting to see reef/skylands data compared to the "older" maps, as I've heard mixed reviews of the map from ladder players. Of course, I've also heard mix reviews about UFO, for example.

That being said, I do agree with you that these provide an argument for suitable ladder maps. Both have average distributions compared to "true and tested" maps in terms of good/ok/bad games, and both have around average minimum/maximum/average game length times.

That being said, it's important to note that this is overall average. Of course it won't reflect individual skill or preference on maps.

Anyway, a few things I want to do:

- See data on Season 5 concerning the updated Skylands and of course ball_cutesy

- Plane data. I'd like to see if I can find correlation of certain plane setups with certain maps.

[Aki1024]

Originally Posted by soccernamlak

- See data on Season 5 concerning the updated Skylands and of course ball_cutesy

- Plane data. I'd like to see if I can find correlation of certain plane setups with certain maps.

Just to make sure you know these pages exist, stats about maps and stats about Skylands (each other map is linked from the first page).

[soccernamlak]

Originally Posted by Aki1024

Just to make sure you know these pages exist, stats about maps and stats about Skylands (each other map is linked from the first page).

Ooooo, ty!

Any data from last season?