Hi all

Vassal Season 5 ended recently. A big thanks to all those organizing it. If you missed it: try to join next time, it is simply awesome!

For those that are not familiar: there are 5 tiers, from Unknown Reaches to Coruscant, with decreasing number of players. These tiers are divided into divisions (named after planets/systems). Every player should play 8 games over the course of several weeks. The best of each division advances, the last goes one back. See here for all you want to know!

One part I also like is how easily available the data is to have a closer look. Unfortunately not all lists are entered. But even without the amount of games is far larger than anything else we have elsewhere. Keep in mind though that many people also like to play a game with less intention to win. This should not be a problem, however, because of the large number. Another problem is that wave 12/13 dropped during the end of the league, which means that there is a small influence of the changing meta visible.

Out of the 1454 games played I could analyze 955 where both players had their list entered. You can estimate the popularity of the factions here:

Those 955 games had a total of 5374 ships on the tables. It looks at first glance af if there were many more rebel ships, but keep in mind that they were played more often. Especially in the unknown reaches:

That is an average 2.81 per player per game. But it is more interesting to look at it per faction and per tier (see the last column for an overall view):

It is interesting to see that the average squadron size increased again compared to the rest of 2017. Personally I'm surprised that imperials are so low, relatively speaking. And whoever dominated the deep core (Coruscant) with scum liked high numbers of ships!

Pilot skill
For pilot skill we have what we might expect:

Please note that these are not percentages but numbers. So yes, there were almost no PS6 or PS11 rebel ships (personally I lost to a PS11 rebel list, so they do exist!).
Imperials are clearly the highPS faction.
Scum seems to be allover the place with PS3/6/9, while Rebels are dominating PS3, 5, 7 and 8. The obvious culprits of Wookies, Jess, Biggs, Lowhhrick, Dash and Miranda are a big part of this, as can be expected.

Generics vs Uniques

There were 1910 squads total. And 1696 squads used at least one unique pilot (88.8%)

That leaves 773 ships in 214 squads, giving an average of 3.6 ships per pure-generic squad.
Those 773 ships are split in the following way:

1. 207 Auzitucks
2. 112 Scum scurrgs
3. 66 Jumpmasters
4. 63 GUNBOATs
5. 56 Rebel Z95s (was that Blair?)
6. 51 TIE Aggressors
7. 34 TIE Strikers
8. 28 A-WINGS!!!
9. 25 Khiraxz
10. 24 Rebel Y-Wings
11. both 15 Scum Z95s and 15 B-Wings

That's 696 of the 773 ships and includes all with more than 10 ships.

Win percentages, overall
But of course that does not tell us yet who won. The next graph shows the win percentages. I had to remove the mirror matches for this, because how would we count an imperial winning against an imperial?

Interestingly enough it looks like Rebels dominate (edit: the other factions in ) the lower tiers and the highest. But inbetween they are figured out by their opponents and the rebel player can't counter the counterplay. Or at least that's one possible explanation.

Arcs
So how about arcs?
That's a bit more tricky. I decided to group it the following way:
Normal arcs are all ships with normal arc + ships with rear arc.
Non-normal arcs are all ships with auxiliary firing arc (180°) + mobile firing arc + primary weapon turrets + the possibility to equip secondary weapon turrets

My reasoning is two-fold: first, a rear arc requires some more skill to get an enemy into arc than all other non-normal arcs. Second, Quickdraw is a massive ship for the Empire. Including rear arcs means that almost every imperial list contains at least a ship with a non-normal arc.
Also, the following two graphs show the win percentage of squads that had at least one non-normal arc, or completely belongs to the normal arc group as I explained (including rear).

Just counting them: there were 1337 squads out of 1910 using at least one ship with a non-normal arc. That is exactly 70.0% and still very high!

And the plot for normal arcs.

Pilot win percentage
So how about win percentages for individual pilots?
Here I decided to throw out all those that played less than 20 games. The imperial trainee for example was used 37 times, but in only 10 games (won 7, lost 3). I think that clears up the picture a bit and leaves us with the following 61 pilots:

It's very interesting to see that Soontir did pretty well, for example. Most are interesting, so please feel free to share thoughts you have when looking at it.

Networks
The last part are my networks again. It's too small to read for pilots, and many are overlapping. In short, I can't handle this massive data size with it. But ships are fine, so I'll show them here. The pilots can be found in the imgur album linked in the end.

Rebels:

Imperials:

Scum:

If you notice anything that seems off, please let me know. Chances are that I missed an NA value or something, but it should be working pretty well by now.
All the analysis was done automatically by using R, which means I can easily rerun it with different arcs, or check pilot win percentages by request.
Many thanks to @sozin for listjuggler, and all the vassal league organizers like @Mu0n729 , @MysteriousRacerX , @VanderLegion and the people I forgot (I'll add them if you let me know!).

The album

Edit: added number of squads with non-normal arc. Also, I had uploaded the wrong plot for winpercentages and arcs. The one I uploaded did not count SWTs as turrets, which is also interesting, but I didn't want that one
Edit: added Unique vs Generics

Also I want to add the following:

I can do something similar for every tournament where you enter all your lists. But I don't do it unless there are enough lists out of the total entered. So please motivate others in your tournament and/or the TOs and judges to enter them into list juggler. Thanks!

Now this is the sort of analysis I love to see. Well done!

A graph I'd love to see would be to go with your normal vs non-standard arcs (and I think you made the right call on rear arcs; only the Bounty Hunter and Phantom II have ones that are full-fat as it is, while ARC & TIE/SF are both "Well, we have one, but..." deals) would be the percentage of squads that took one.

Which is to say, how often when setting up across the table were you expecting to see a turret-alike firing arc? That can be useful for tracking how often you face one, regardless of how well it does.

It's hilarious to me how the entire Imperial faction revolves around Quickdraw.

28 minutes ago, Favoritism Flight Games said:

It's hilarious to me how the entire Imperial faction revolves around Quickdraw.

I'm actually becoming increasingly convinced that Quickdraw is completely broken, but somehow each wave we end up with something so much worse that he keeps flying under the radar.

34 minutes ago, DR4CO said:

I'm actually becoming increasingly convinced that Quickdraw is completely broken, but somehow each wave we end up with something so much worse that he keeps flying under the radar.

It's very difficult for a thing with limited arcs, no regen, and modest defense to really be broken.

But she is awfully good.

And perhaps more importantly, she's good against things that are broken.

1 hour ago, Biophysical said:

It's very difficult for a thing with limited arcs, no regen, and modest defense to really be broken.

But she is awfully good.

And perhaps more importantly, she's good against things that are broken.

It makes you wonder if Dengar will be back for Scum since he fits a similar role to QD (albeit for more points but with more health and a bit faster due to big base)

WAIT.

There’s 66 data points for YV-666?

?

Please note that I had uploaded the wrong plots for arcs. I also had a version where SWTs are not counted as non-normals because they are not necessarily equiped. This means now that there were almost no squads for Scum that won if they had just normal arcs.

5 hours ago, Reiver said:

Now this is the sort of analysis I love to see. Well done!

A graph I'd love to see would be to go with your normal vs non-standard arcs (and I think you made the right call on rear arcs; only the Bounty Hunter and Phantom II have ones that are full-fat as it is, while ARC & TIE/SF are both "Well, we have one, but..." deals) would be the percentage of squads that took one.

Which is to say, how often when setting up across the table were you expecting to see a turret-alike firing arc? That can be useful for tracking how often you face one, regardless of how well it does.

Thanks!

There's not much of a graph for that as it boils down to those two numbers: 1337 squads out of 1910 took at least one, that's exactly 70%. 573 used only normal arcs.

23 minutes ago, Ailowynn said:

WAIT.

There’s 66 data points for YV-666?

?

66 ships out of 5374 were YV666s, yes.

One more thing: Generics vs Uniques

There were 1910 squads total. And 1696 squads used at least one unique pilot (88.8%)

That leaves 773 ships in 214 squads, giving an average of 3.6 ships per pure-generic squad.
Those 773 ships are split in the following way:

1. 207 Auzitucks
2. 112 Scum scurrgs
3. 66 Jumpmasters
4. 63 GUNBOATs
5. 56 Rebel Z95s (was that Blair?)
6. 51 TIE Aggressors
7. 34 TIE Strikers
8. 28 A-WINGS!!!
9. 25 Khiraxz
10. 24 Rebel Y-Wings
11. both 15 Scum Z95s and 15 B-Wings

That's 696 of the 773 ships and includes all with more than 10 ships.
(to be continued)

Generics continued:

Win percentage for generics

You'll note the similar numbers, but they can be explained with the amount of games played, which is very low:

And of course squad size is way above the average for the rest, which is naturally expected as generics are also way cheaper, in general:

Also, no generics for imperials in the deep core, and no imperial generics winning in core worlds.

@GreenDragoon Thanks for running this data!!! Would you be willing to re-run this plot for games run after the "great nerfing" only?

31 minutes ago, acidReign said:

@GreenDragoon Thanks for running this data!!! Would you be willing to re-run this plot for games run after the "great nerfing" only?

Sure, all of it should not be a problem. If I recall correctly, there is a column with the date in the raw data. If that‘s not empty then the subset is trivial

Thanks that would be awesome!

Nice work @GreenDragoon ! Engendering this sort of scientific analysis is what caused me to create List Juggler in the first place. Thank you!

3 hours ago, acidReign said:

Quickdraw = Highest use, but win % is 50%.

If you consider that (besides the unkown new league members placed into the Unkown Reaches) the continuity of the league has created groups of similarly skilled people (where we can make a modest assumption that each match is played at a roughly similar skill level), then Quickdraws performance as a whole makes her an Average pilot in the grand scheme of things.

And Lowhhrick and Miranda look awesome/OP/easier mode.

Of course, XWing is a squad game, so you have to account for the full squad. QD has a million Tier 2.0 partners, and even a large # of Tier 1.0 partners, so squads can vary greatly (and strength vs matchups varies greatly). Miranda and Low are currently pretty married to each other or some similar strong partners, and people don't really deviate a whole lot, so the result is more consistent wins across the board.

6 minutes ago, phild0 said:

Quickdraw = Highest use, but win % is 50%.

And Lowhhrick and Miranda look awesome/OP/easier mode.

This, the summary @phild0 .

Well done @GreenDragoon , you get the Star Geek of the day award!

1 hour ago, acidReign said:

Thanks that would be awesome!

Simple I said... I need to convert the "2017-10-28 08:30:00 CEST" factors into dates etc. So it's not quite as simple... I'll probably get to it tomorrow though. Edit: Can't stop now, can I? Bad news is that only 239 games with both lists entered were played after. But then again, that's more than most tournaments, so maybe it's still useful

1 hour ago, sozin said:

Nice work @GreenDragoon ! Engendering this sort of scientific analysis is what caused me to create List Juggler in the first place. Thank you!

Thanks, would not be possible without your work though!

30 minutes ago, phild0 said:

Of course, XWing is a squad game, so you have to account for the full squad.

Which reminds me: I did not mention it, but the winrates for pilots do actually include games against a squad where a mirror happened. Tracking all opponents through the whole process is something I can't do. I can and did filter for mirror matches with respect to factions, but that information is not lost. The exact list of the opponent is lost during the process. (But writing this gave me an idea how I could actually get it back, so maybe I will do it eventually. There will be a season 6, after all)

It does mean that very frequently used ships as Quickdraw will have a bias towards 50% because they are more likely to face each other, where the average is of course 50% for the two identical ships facing each other. But we don't know whether the ship would have won or lost more without the mirror matches, so we can't say that the actual values are a bit higher.

21 minutes ago, clanofwolves said:

you get the Star Geek of the day award!

Thanks, I'll take it

This is great stuff. Thank you for all the work you did.

9 minutes ago, phild0 said:

Quickdraw = Highest use, but win % is 50%.

If you consider that (besides the unkown new league members placed into the Unkown Reaches) the continuity of the league has created groups of similarly skilled people (where we can make a modest assumption that each match is played at a roughly similar skill level), then Quickdraws performance as a whole makes her an Average pilot in the grand scheme of things.

And Lowhhrick and Miranda look awesome/OP/easier mode.

Of course, XWing is a squad game, so you have to account for the full squad. QD has a million Tier 2.0 partners, and even a large # of Tier 1.0 partners, so squads can vary greatly (and strength vs matchups varies greatly). Miranda and Low are currently pretty married to each other or some similar strong partners, and people don't really deviate a whole lot, so the result is more consistent wins across the board.

The alternative explanation is that Quickdraw's floor is a lot lower compared to Miranda's ilk. So someone not knowing what they're doing still does okay with Miranda, but gets Quickdraw quickly tanked. This is sort of what you said, but maybe a little different.

1 hour ago, acidReign said:

Thanks that would be awesome!

Here you go:

Win percentage for pilots based on 239 games after the FAQening on November 6th. Minimum of 10 games used as threshold. This leaves 34 pilots.

The identical/overlapping ones are Omegaleader&Lokrevenant at 40%, Talonbane & Ketsu at 45%, Ryad&Omicron at 50%, Wedge & Dengar at 66.6%.

Also interesting how Quickdraw fell to 42.6%, while Miranda and Low fell less (55.3% and 54.7%, respectively) but swapped places.

Bonus edit:

The networks for this subset

I think it's fascinating how wookies are really dominating now! Also: 19 Greensquadrons?! Maybe that points at logged games where someone was through with his 8 and played some more purely for fun?
Imperials are still QD, Vader, Inq, RAC and now also Rho squadron
Scum is Assajj, Fenn and Lokrevenant

That Rho squadron is showing up at all is nifty.

There haven't been many Imperial generics showing up at all, so it's pretty exciting.

Hey! I got a good request for you. My local group has a few players (justifiably) starting to complain about harpoon missiles - how have they helped the Vassal season 5 players to win across the tiers?

6 hours ago, Mu0n729 said:

Hey! I got a good request for you. My local group has a few players (justifiably) starting to complain about harpoon missiles - how have they helped the Vassal season 5 players to win across the tiers?

[Outer Rim] Anecdotaly the first times I used them I blew away my opponents, but as the league progressed I had worse results. Some of them were against other harpoons, though.

6 hours ago, Mu0n729 said:

Hey! I got a good request for you. My local group has a few players (justifiably) starting to complain about harpoon missiles - how have they helped the Vassal season 5 players to win across the tiers?

Anecdotally, as well, I never used them.