11 coaches online • Server time: 05:11
Index Search Usergroups Profile
Log in
Recent Forum Topics goto Post Replaysgoto Post Should I fire -MAx2 ...goto Post FUMBBL vlog
SearchSearch 
Post new topic   Reply to topic
View previous topic Log in to check your private messages View next topic
Sp00keh



Joined: Dec 06, 2011

Post   Posted: Jun 12, 2018 - 23:30 Reply with quote Back to top

I created a PowerBI dashboard looking at the winrates, casuality rates, etc for different races in FUMBBL

I spent ages on this, but didn't really feel it was finished so kind of left it alone for a long time
Now though there's been a bunch of questions on forum about which races are good at stuff, so I've dug it back out and publishing anyway

- There's 11 pages to it, please click through
- It's interactive, you can click on stuff like column headers to Sort, Filter etc
- Some of the pages have green or black squares, you can use these to filter by B or R

https://app.powerbi.com/view?r=...

- It's data from FUMBBL matches in Blackbox and Ranked
- Date period spans march to november last year, starting after the time when fumbbl implemented BB2016 rules eg removed Piling On skill
- 72k games

I hope you like it
sann0638



Joined: Aug 09, 2010

Post   Posted: Jun 12, 2018 - 23:37 Reply with quote Back to top

Awesome - have you seen my similar ones for NAF data in Tableau? https://public.tableau.com/profile/mike.sann0638.davies#!/

_________________
Image
NAF Ex-President
Founder of Cakebowl, Wiltshire's BB League @ cakebowl.co.uk
Magicporncup



Joined: Sep 30, 2015

Post   Posted: Jun 13, 2018 - 00:12 Reply with quote Back to top

one word : amazing
Klazam



Joined: Aug 01, 2015

Post   Posted: Jun 13, 2018 - 00:16 Reply with quote Back to top

Wow.

Takeaway- Ogres need mucho help

and this game is suprisingly well-balanced
awambawamb



Joined: Feb 17, 2008

Post   Posted: Jun 13, 2018 - 00:54 Reply with quote Back to top

Flings cause almost twice the pain of pro elves.
WAT

_________________
"la virtú sta nel cielo e nella terra, ma anche nelle nuvole e nelle stelle"

Image
gpope



Joined: Jun 04, 2010

Post   Posted: Jun 13, 2018 - 01:22 Reply with quote Back to top

awambawamb wrote:
Flings cause almost twice the pain of pro elves.
WAT


More Mighty Blow, more fouls, more kickoffs received, fewer bash turns cut short by scoring.
Markusen



Joined: Jan 23, 2005

Post   Posted: Jun 13, 2018 - 02:51 Reply with quote Back to top

Poor Ogres, poor Goblins...
Can't score, can't bash...
Just some SPP delivery boyz.
mrt1212



Joined: Feb 26, 2013

Post   Posted: Jun 13, 2018 - 04:22 Reply with quote Back to top

Sp00keh wrote:
I created a PowerBI dashboard looking at the winrates, casuality rates, etc for different races in FUMBBL

I spent ages on this, but didn't really feel it was finished so kind of left it alone for a long time
Now though there's been a bunch of questions on forum about which races are good at stuff, so I've dug it back out and publishing anyway

- There's 11 pages to it, please click through
- It's interactive, you can click on stuff like column headers to Sort, Filter etc
- Some of the pages have green or black squares, you can use these to filter by B or R

https://app.powerbi.com/view?r=...

- It's data from FUMBBL matches in Blackbox and Ranked
- Date period spans march to november last year, starting after the time when fumbbl implemented BB2016 rules eg removed Piling On skill
- 72k games

I hope you like it


Any chance we could expand the date range at some point?
Sp00keh



Joined: Dec 06, 2011

Post   Posted: Jun 13, 2018 - 08:57 Reply with quote Back to top

Date range expansion is pending

Ogres are shocking at everything, apart from they are good at inflicting CAS.
Christer



Joined: Aug 02, 2003

Post   Posted: Jun 13, 2018 - 10:52
FUMBBL Staff
Reply with quote Back to top

I love this kind of thing. Good job Smile

(Also edited the display of the long URL to avoid the layout break)
PurpleChest



Joined: Oct 25, 2003

Post   Posted: Jun 13, 2018 - 10:56 Reply with quote Back to top

Beautiful numbers and pretty number arrangements. Awesome work.

I find the data on page 5 to be presented unhelpfully as a linear graph.

The instant big take away for me is the close correlation between td's and win rate. Bashings impact on win rate seems almost negligible. but i look forward to spending some time drinking these numbers in more.

_________________
Barbarus hic ego sum quia non intelligor ulli
I am a barbarian here because i am not understood by anyone
Christer



Joined: Aug 02, 2003

Post   Posted: Jun 13, 2018 - 11:41
FUMBBL Staff
Reply with quote Back to top

If we're doing requests, add TV difference and CR difference as parameters for the analysis Smile
Chivite



Joined: Sep 04, 2017

Post   Posted: Jun 13, 2018 - 12:07 Reply with quote Back to top

wow, this is some impressive work.

it truly shines a new light on the experience i already had of the game
Christer



Joined: Aug 02, 2003

Post   Posted: Jun 13, 2018 - 13:15
FUMBBL Staff
Reply with quote Back to top

I do similar types of extracts manually from time to time, so the data doesn't surprise me at all. I haven't taken the time to set up something like PowerBI or Tableau though, nor have I taken the time to build an OLAP cube for it all.. Would certainly be super interesting, but oh so time consuming to find suitable tech for it.

At work, I primarily work with SSAS (Microsoft SQL Server Analysis Services), but that's quite a bit beyond being reasonably priced for FUMBBL, so I'd need to figure out the whole technical platform (Apache Druid, Mondrian, MariaDB Column store, probably others available out there) and do all the work to set it up.. Maybe some day Smile
Sp00keh



Joined: Dec 06, 2011

Post   Posted: Jun 13, 2018 - 13:48 Reply with quote Back to top

Power BI can connect directly to SQL Server, for live data reports. Also data cubes etc, all kind of things
This uses a CSV that koadah extracted for me, I've asked for an update but maybe Christer can help? Smile

I do want requests
It doesn't have coach rating in the CSV
I could estimate coach skill from their observed winrate in the sample but it's only an average of 31.5 games per coach so is dubious to do so
I'd be interested to see if there's a correlation between coaching skill and races played tho


Adding TV difference is possible but makes it harder to GroupBy.. I'm not sure which visualisation it's appropriate for?
Display posts from previous:     
 Jump to:   
All times are GMT + 1 Hour
Post new topic   Reply to topic
View previous topic Log in to check your private messages View next topic