There are still a few more Regionals to come in Canada, Europe, and Australia. I will add those results as well when they come out, but probably not til the end of July as I will be out traveling while they are going on.
If you haven't looked over the data yet, its HERE
This as been a really interesting project. It got started when someone (brikhause on the FFG forums) posted the full fleet lists and standings for the San Antonio Regional. It got me thinking that with that kind of information you could start doing some comparisons to see if there were any trends in the fleets showing up on the top tables and the general population. I am a science teacher, and hence a bit of a data nerd, so this seemed like a fun project.
What am I trying to answer?
The primary question I was trying to answer was: "Is there a difference between the fleets showing up on the top tables and the general attendance of the regional tournaments?"
Is the data any good at all?
There are definitely limitations to the data and what it can tell us. With data from 27 different regionals, the sample size is pretty good. I have data on a good number of fleet types and enough to start to see some patterns. As time went on and a new tournament was added to the data set, the numbers only adjusted slightly, which gave me confidence that a pattern was forming. So where are the limitations?
- While I know the winners of 27 of the tournaments, I only have the full fleet lists for 8 of the tournaments. That does include 143 total fleets, which is nice, but we have to rely on those 8 tournaments to give us the big picture snapshot. It would be nice to have more tournaments with full data available.
- Obviously, final tournament placing is about far more than just fleet composition. Luck in dice and pairings as well as player skill make up the majority of what goes into winning. To that end, the data on the Top 4 or Top 8 becomes a little more interesting as most of the players finishing near the top should be fairly skilled. Then we run into the limitation of having fewer tournaments to draw from.
- As I had to rely on lots and lots of help to gather the data there wasn't a really consistent method in place for doing so.
So, what is it good for?
Well, there are definitely things it is not good for. You can't look at the data and use it to build the ultimate, unbeatable fleet. While a Motti-led, Rhymerball, with Demolisher has done pretty well, it has lost plenty of times as well.
What the data is good for is giving you an idea of what you might expect to run into. Even this can vary by area, but it does give you a general idea. You can look at the data and have an idea of the types of fleets you might see on the top tables and consider if you are prepared for them.
It also says a little about the overall balance of the game or the prevalence of various ships and upgrades. Some patterns popped up on what ships, admirals, and upgrades were seeing the most play as well as which were showing up more frequently on the top tables. I'm not sure that this is so much useful data as just really interesting. At least it is to me, but we already established that I am a data-nerd.
What did we learn?
Let's take it one section at a time
Rebels vs Imperials
Over the span of this project, this is probably the data that changed the most and was sometimes the most alarming. From start to finish more Imperials have been showing up at the tournaments than Rebels. The gap was wider at times, but really seemed to settle at the 55%/45% fleet favoring the Imperials. There are plenty of reasons why that might be and I think one is certainly just that Star Destroyers look really cool.
What was alarming was the rate that the Imperials were dominating the top tables. At one point they were winning 80% of the tournaments and made up 75% of the Top 4. That spoke to a potential for game imbalance. Over time that largely settled down and the data for the different groups settled down to be similar to the overall attendance. With the exception of the tournament winners. The Imperials ended up winning about 2/3rds of the Regionals tournaments that we know of. Given the similarity of the other groupings, it certainly stands out. If the Top 4 were also skewed to the Imperials, I would be more concerned about it. Either way, it is an interesting data point.
This is the first bit of really crunchy data we have and a source of surprise to me. I can honestly say that I did not take Motti very seriously before this project. I've found him rather bland. Motti was not only taken very frequently, he won at an even higher rate. Anecdotally, Motti seemed to be the Admiral of choice in Imperial fleets based around Rhymer, which was also very successful.
Otherwise, this data establishes the Big 4 Admirals for the 2016 Regional season: Ackbar, Rieekan, Motti, and Screed were taken more than other Admirals and saw a large amount of success as well. Dodonna, Mothma, Garm, and Ozzel all saw a small amount of play and some success although they didn't pull out a win. I think that is more due to how often they were taken than their potency.
The real outliers in this data are Tarkin and Vader. Tarkin showed up about as often as the other second tier Admirals, but had almost no success. I think that their are some interesting tricks with Tarkin and was pretty surprised at this. Maybe his cost makes more competitive players avoid him? Vader just barely showed up at all. Almost all of his success can be attributed to one player (MikeMcMann on the FFG boards). This again surprises me quite a bit. I really like Vader and think he can be really effective. He may eventually find a home if the Imperials gain some ships with lots of red dice.
Interesting note is that the 2 most popular Rebel Admirals were their most expensive, while the 3 least expensive Imperial Admirals were the most popular.
The Archetype data is possibly more interesting than useful. Part of that being it is the most subjective of the data in the set. Initially I had quite a few more categories than the 9 that I settled on. While the increased granularity was nice, it made it too hard to see patterns in the data. If I do this process again, I will certainly solicit some help in establishing useful categories.
That being said, the data still is pretty interesting. What obviously stands out is just how dominate fleets based on Major Rhymer were in this Regional season. If anyone can be said to have won the Regionals this year, it was Rhymer.It was the most common archetype in both the Top 4 and overall winners. Thanks to Rhymer, Imperial fleets built around squadrons generally outperformed their Rebel counterparts.
Another bit that stands out were the non-Demolisher Imperial swarms. There weren't many of them, but those that did show up outperformed their prevalence. Rebel Swarms did both well and poorly. They were the 3rd most winning archetype, but showed up in the bottom 25% more frequently than any other. This actually matches my personal experience running a Mothma Swarm. I either win big or lose big, it seems to be a feast-or-famine archetype.
This was a data set that I was particularly interested in. My own bias had been telling me that high activation fleets were the way to go in Armada and they would out perform lower activation fleets in Regional tournaments. I think it is safe to say that the data does not support that assertion. Another case where my assumptions were not supported by the data.
Overall, the data on fleet size is pretty much just noise with no obvious trend. The number of squadrons and deployments in a fleet do seem to show a correlation with success. In both cases, as you moved up to the top tables, the average number of squadrons and deployments increases steadily. It's not a huge amount by any means, but the pattern still exists.
Where this points me is to the somewhat underrated value of deployments and having a strong deployment strategy. It's something I have begun to spend more thought on in my game play.
Ship Class Prevalence vs Performance
This is another interesting data set, that goes against a bit of what I was expecting, while confirming some other suspicions. Demolisher Gladiators are generally considered one of the more effective ships in the game. Not surprisingly, they were among the most common ships brought and the most common ships in winning fleets.
What was a bit surprising was the somewhat poor relative performance of the MC-30. I like them and consider them a really solid ship, but they didn't show up near the top of tournaments quite as often as they were brought. In some ways, the Neb-B and VSD come out looking a bit better than their reputations.
|All||Bottom 1/4||Top 1/2||Top 8||Top 4||Winners|
I only tracked a few specific upgrades and generally those that were most interesting to me. I didn't originally track Yavaris, but added it based on a suggestion. There are some interesting trends to look at here:
- Demolisher - Easily the most common title in the game and it wasn't until late in the Regional season that a fleet showed up that had a Gladiator without having the Demolisher title in it. It generally outperformed its expectations and showed that it deserves its reputation as the best upgrade in the game. Almost half of the winning fleets in all the tournaments had Demolisher, which is a bit nuts.
- Rhymer though had quite an impressive showing as well. He didn't show up often, but had a large share of success.
- XI7s - I looked at XI7s not because I felt they were too strong, but more at their impact on Advanced Projectors. Not many took APs in their fleets, because of the fear that XI7s make them irrelevant. With XI7s showing up from a quarter to a third of the time it is hard ot know if APs are a good choice. But the data may help you decide.
- TRCs - I think TRCs are among the strongest upgrades in the game. It is some really powerful manipulation of the dice that need it the most. The way they play, makes them almost an exclusively Rebel card and they seemed to perform pretty well for the Rebel fleets that took them.
- Yavaris - I added in Yavaris when someone compared it to Demolisher, saying it was nearly as powerful and that no one took a Neb-B without it. Well, the data doesn't back that up as Neb-Bs seemed to show up about 3x more often than Yavaris.
- 0 Squads - I tracked this out of curiosity. I just wanted to see how well no squad lists would fair in the Wave 2 environment. Not great I guess. They didn't show up very often, and when they did, performed worse that their prevalence. People did win with no squadrons, so it can work, but it certainly goes against the trend.
I gathered a little bit of other data as well. It looks like the winners were scoring 8-9 VPs per round regardless of the size of the tournament. I grabbed a little data on bids, but it wasn't very instructive as the sample size was just too small.
So, was it worth it?
I think so. It looks like I answered my question, there are some trends in the differences between the overall attendance and the top tables. I also learned that I had some misconceptions in my assessment of what would do well in the tournaments.
Will I do it again?
I think so. It was fun putting together the data and seeing what patterns emerged. Barring life being too hectic, I will likely compile the results for the 2017 Regionals season as well.
What would be different?
I'd like more community input on data collection and analysis. I think I could get some help setting up the archetypes to track. I'd like to make gathering the data a little easier so that we can get the full fleet lists for a higher percentage of the tournaments.
So, please let me know what you think of this project. Was it useful? Did anything stick out to you in the data? Was I wasting my time?