ladle-export: analysis of results, teams, and players

[dlh]

I wrote a program to export all the past Ladle results to JSON. You can find the code on Launchpad at Ladle Tools. With this data you can generate stats and other reports.

I made an implementation of Concord's idea for Ladle stats. Source, Live demo

The other branch is a program to print out stats about players—Who won the most Ladles? What Ladles has this person played in? Who has played in the most Ladles? Examples:

Code: Select all

$ python ladle-results.py aliases.json ../ladle-export/ladles.json report_wins
6  newbie         Ladle 4, Ladle 6, Ladle 9, Ladle 11, Ladle 21, Ladle 24
6  pavelo         Ladle 6, Ladle 9, Ladle 11, Ladle 14, Ladle 18, Ladle 19
...

Code: Select all

$ python ladle-results.py aliases.json ../ladle-export/ladles.json report_occurrences
21  sasha          Ladle 1, Ladle 3, Ladle 4, Ladle 5, ...
19  g5             Ladle 1, Ladle 2, Ladle 3, Ladle 4, Ladle 7, ...
...

Code: Select all

$ python ladle-results.py aliases.json ../ladle-export/ladles.json report_player freekill
Results for freekill:
	Ladle 28 with The Usual Suspects (Lost in Finals)
    ...

A few Ladle pages have no player listings, so if you feel like doing some bookkeeping go through the Past Ladles page and try to fill out the missing details.

[INW]

im guessing this is for help with my website and abilities to get previous ladle statistics?

if it is, thats great, but idk how to use the stuff, someone will have to teach me (:

------>viewtopic.php?f=2&t=19821

[epsy]

look at the source he linked

[Lackadaisical]

I see the quality score needs to be revised i'm pretty sure the runner ups (in this case) shouldn't have had a higher quality score than winner?

edit: It seems the calculation is just wrong for tx, it should be

2*0(vcl match) + 2*1(ct beta) + 2*2(team pl) + 2*3.111(the usuals) = 12.222

[dlh]

I see the quality score needs to be revised i'm pretty sure the runner ups (in this case) shouldn't have had a higher quality score than winner?

edit: It seems the calculation is just wrong for tx, it should be

2*0(vcl match) + 2*1(ct beta) + 2*2(team pl) + 2*3.111(the usuals) = 12.222

Yes, it was a bug. It has been fixed. The final quality score is still different than Concord's version—see the source code for the difference.

[dlh]

If you're interested in analyzing Ladle results, then you can use the exported JSON data available from http://generalconsumption.org/armagetro ... le-export/.

A while back I added Team stats to ladle-results (this program uses the exported Ladle data):

Code: Select all

$ python ladle-results.py ../ladle-export/ladles.json info "Rogue Tronners"
Results for Rogue Tronners:

             Wins  Losses  Played  % Won 
    Matches   87     51     138    63.0% 
    Bouts     39     20      59    66.1% 
    Ladles    2      20      22     9.1% 

            Achieved  Missed  Chances  % Achieved 
    Seeds      13       5        18      72.2%    
    Semis      14       8        22      63.6%    
    Finals     5        17       22      22.7%    
                                                         
    Ladle 23  _~Rogue Tronners~_  Lost in Round of 16    
    Ladle 34  Rogue Tronners      Lost in Quarter Finals 
    Ladle 35  Rogue Tronners      Lost in Quarter Finals 
    Ladle 37  Rogue Tronners      Lost in Semi Finals    
    Ladle 38  Rogue Tronners      Won Tournament         
    Ladle 39  Rogue Tronners      Lost in Finals         
    Ladle 40  Rogue Tronners      Lost in Quarter Finals 
    Ladle 41  Rogue Tronners      Lost in Semi Finals    
    Ladle 42  Rogue Tronners      Lost in Round of 16    
    Ladle 43  Rogue Tronners      Won Tournament         
    Ladle 44  Rogue Tronners      Lost in Semi Finals    
    Ladle 45  Rogue Tronners      Lost in Finals         
    Ladle 46  Rogue Tronners      Lost in Round of 16    
    Ladle 48  _~Rogue Tronners~_  Lost in Finals         
    Ladle 49  _~Rogue Tronners~_  Lost in Semi Finals    
    Ladle 50  _~Rogue Tronners~_  Lost in Quarter Finals 
    Ladle 51  _~Rogue Tronners    Lost in Semi Finals    
    Ladle 52  Rogue Tronners      Lost in Semi Finals    
    Ladle 54  Rogue Tronners      Lost in Quarter Finals 
    Ladle 55  _~Rogue Tronners~_  Lost in Semi Finals    
    Ladle 56  _~Rogue Tronners~_  Lost in Semi Finals    
    Ladle 57  Rogue Tronners      Lost in Semi Finals    

I know sine.wav is tracking some of these stats manually, and hopefully this data can save him (and others) some work in the future.

Edit: The term “round” is ambiguous in the context of Armagetron, so I use “bout” to describe a round in the bracket where two teams play matches.

[PokeMaster]

Looks cool, but I don't know how to do poop with it

[sinewav]

Edit: The term “round” is ambiguous in the context of Armagetron, so I use “bout” to describe a round in the bracket where two teams play matches.

I prefer the term bout as well. Concord says game. I'm recording advances through the bracket as rounds to go along with "opening round."

I really don't want to record stats because it's a huge pain in the arse. Yours will clearly be more accurate than mine anyway, missing human error. What I'd like to see is the stats become useful and meaningful in the future of Ladle. That's why I'm going to push this yearly tournament idea. Once we have a season and a playoff we should abandon the wiki stats (except for the very useful server tracker) and redirect to generalconsumption.

I'm going to bump the SuperLadle thread and cross-link here because I think the conversation needs to continue there.

[Concord]

I do say game, but I don't like it. bout is better.


I don't like it because Tennis uses game and match as well, but the meanings are the opposite of ours.