Anyways- to tie this back to my point... I'd love to see it get granular so we can enable analysis. To do that the data needs to be atomic... on the other hand, that is asking a LOT for TOs. I'd love to have a game by game breakdown!
Game by game would be awesome and would be the best way to really look at Objectives. Given how challenging it was to even get full fleet lists last time around, that may be impractical.
I will at least grab it at the tournament that I TO.
I agree. I think it's sadly too impractical. You could still add the objectives to the lists, and draw SOME conclusions there... I did build a massive "After Action Report" community effort with Battlefront Miniatures many years ago. Had over 2000 responses that offered some really good insights.
Ergh, data collection for data collection sake is meaningless.
This is not so true these days! Data Lakes and Big Data say- gather it all and figure stuff out later!
They are actually starting to say the opposite. Realised that gathering large amounts of data is essentially pointless as it breaks the scientific process and then becomes too overwhelming to be of any use.
As a scientist, I disagree. Gathering large amounts of data is extremely useful in identifying trends and other global movements. However, it is not very useful when looking at precision or accuracy because everything blends together. I enjoy both aspects of data analysis, but you need a specific purpose for that set of data.
Edit: I'd like to add this data is trending because it is collected over a long period of time at different intervals so we should collect as much as possible.
Edited by Undeadguy
Ergh, data collection for data collection sake is meaningless.
This is not so true these days! Data Lakes and Big Data say- gather it all and figure stuff out later!
They are actually starting to say the opposite. Realised that gathering large amounts of data is essentially pointless as it breaks the scientific process and then becomes too overwhelming to be of any use.
I would say it depends on the amount of data and the project. In our case its about capturing and storing actual data as opposed to processing it and storing only results. As we don't know what questions will be interesting/useful by the middle of the regional season (did anyone envision Riiekan Zombies as a powerful archetype at the beginning of wave 2?) it's better to have initial data available.
Ideally, I would capture the following (thats what I did in one of the regionals):
- All fleet lists, or at least top8/top16 (just ask organizer and take a picture of every list with your phone)
- Tournament standings at the end of the tournament
- Pairings for every round (this helps to find out what archetypes were encountered/defeated by top lists. Also it can help to uncover the contents of the top list that you missed by asking one of the opponents)
Edited by pt106Tableau is great but pricey! Periscope could do?
PLEASE no PIE Charts.
Fair point. I had it on free trial, hence why I thought it was free
What can be done with the data depends heavily on collection. It was pretty variable what I got last season and sometimes a challenge to piece it together. Fortunately, the data is at least concrete enough that it was generally trustworthy. Data that can be collected from easy to hard::
It would be nice to move further down the list, but becomes a real chore for whoever is doing the collection. Processing it later takes time, but that work can be dispersed if necessary.
We had data from 27 tournaments last time, but only full fleet lists for 8 of them. It would be great to move that number over 50% this time around.
Any way the community can help with the acquisition of data?
I'm not sure it is feasible to track every upgrade in the game. Last time I tracked:
Demolisher
Rhymer
XI7s (to check the viability of APs)
TRCs
Yavaris
I would say this time I would add:
APTs
BCC
Targeting Scrambler
Jan/Dengar/Intel?
Rogue vs non-rogue
What else?
Any way the community can help with the acquisition of data?
The best is to attend a tournament and grab everything you can. Then either type it up on here or share it with me.
I'm brainstorming ways to make data acquisition easier, but it really relies on having a dedicated volunteer on site.
Any way the community can help with the acquisition of data?
The best is to attend a tournament and grab everything you can. Then either type it up on here or share it with me.
I'm brainstorming ways to make data acquisition easier, but it really relies on having a dedicated volunteer on site.
I'd assume you want date, store location, amount of people, their lists, and the rankings. Any reason why someone would say no to this? Besides RRG stating no outside material references?
Any way the community can help with the acquisition of data?
The best is to attend a tournament and grab everything you can. Then either type it up on here or share it with me.
I'm brainstorming ways to make data acquisition easier, but it really relies on having a dedicated volunteer on site.
I'd assume you want date, store location, amount of people, their lists, and the rankings. Any reason why someone would say no to this? Besides RRG stating no outside material references?
No issues that I know of or that were reported last time. It's not a problem to collect the data, some one just has to be there to do it.
Rogue vs Non Rogue is a good one.
With flotillas and FCT i envisage a massive reduction in number of rogues being run, and a massive increase in number of squads.
Ergh, data collection for data collection sake is meaningless.
We need hypothesis' to test. So I propose one statistic.
Number of B wings in fleet.
Collection of the data should be separate from running analytics on the data. Ideally different people. Collect the most data possible, then let others look for trends.
Woooo! This was a great tool last time around. I would suggest keeping the number of specific upgrades you track limited. Demolisher and Rhymer are both obvious inclusions. I don't think APTs, XI7s, or TRCs are worth tracking this time around. BCC, Targeting Scrambler, and Rogue vs non-Rogue seem like a good idea. I really like the idea of tracking objectives in some form or another.
Edited by TruthinessGlad to see this making a return! The data collected heavily influenced my regionals list and I ended up first as a result... The other half of the influence came from iceqube's discussion of objectives and so I want to second his suggestion here that objectives be tracked, preferably all objectives that were taken, not just the ones chosen... if possible
Edited by SkyCakeRogue vs Non Rogue is a good one.
With flotillas and FCT i envisage a massive reduction in number of rogues being run, and a massive increase in number of squads.
On the surface I agree, but I think 8xYT2400 is still very strong and I posit that it should be tracked.
I'm with Shmitty on his labor of gathering the data. Ultimately it may be best to focus on the core info that the tournament software gathers plus lists. Being too **** specific without a dedicated volunteer on site to do the gathering is asking for incomplete data or ill will from those who arn't information freaks/forum dwellers.
And my initial reaction to the topic was "Tracking how many random people pound me into the ground during my swing at Regionals 2017".
I would definitely be interested to see how many slicer tools make it in
I'd be interested overall in seeing how well represented each wave is.
Between the core set to wave 5 there will be a great deal of variation, but I'd like to see which wave is the clear winner.
I imagine the results will be slightly skewed towards 5 because of shiny, but maybe things will reach some sort of excitement equilibrium by then.
I'd be interested overall in seeing how well represented each wave is.
Between the core set to wave 5 there will be a great deal of variation, but I'd like to see which wave is the clear winner.
I imagine the results will be slightly skewed towards 5 because of shiny, but maybe things will reach some sort of excitement equilibrium by then.
That Data will be skewed itself if we don't have Wave 5 at the Start of November.
As Regionals start with the Vancouver Regionals on the 12th of November ,I believe... So there woul dbe no Wave 5 Representation until they did drop, and 11 days have passed.
It'll have ship breakdowns and it wouldn't be too hard to parse that into waves I guess. Not sure it would have much meaning beyond curiosity for the reasons Dras stated.
Given then timing of the Wave 5 release, the data may end up split.
I have numerous curiosities that I'd like settled and answers...
But really, compared to even just attempting to repeat last time, they are small, insignificant, and wholly selfish 
I have numerous curiosities that I'd like settled and answers...
But really, compared to even just attempting to repeat last time, they are small, insignificant, and wholly selfish
Oh come on... This whole thing started with me wanting to satisfy my idle curiosities. What do you want to know? If it's feasible, I can fit in.
I have numerous curiosities that I'd like settled and answers...
But really, compared to even just attempting to repeat last time, they are small, insignificant, and wholly selfish
Oh come on... This whole thing started with me wanting to satisfy my idle curiosities. What do you want to know? If it's feasible, I can fit in.
Its little things like - All ships have access to the Officer slot, so its which ships are taking faction specific, vs generic officers, to try to get a handle of Skilled First Officers / Liasons / FCs compared to Leias, Raymus' and Wullfs', sort of thing...
Its overly complex for the satisfaction - especially since the greatest challenge to it is the reporting. And reporting would have to be spot on to gather that data as meaningful.
I generally have the statistics and data management on, so I understand what it would take, and still advise away from it 
That actually isn't bad. The goal is for full fleet lists. If we have that it is just another box to check on data entry.
I can even give you edit access and make you enter that data yourself.