6 hours ago, Scott Pilgrim2 said:

Guessing that 7B Obiwan is better than CLT Anakin.

I'd much rather have the consistent +1 attack and +2 health than fishing for that bullseye to do any meaningful damage.

With only 17 people at the Anchorage Hyperspace Trial, I can only imagine how demoralizing it would be to be the one guy at the whole tournament that doesn't walk away with dice.

Hey @catachanninja, you know what I'd love more than that definition of an ace?

A filled out hyperspace trial instead of 27/59

51 minutes ago, GreenDragoon said:

Hey @catachanninja, you know what I'd love more than that definition of an ace?

A filled out hyperspace trial instead of 27/59

It just happened. And he's been harassing to get it to 27.

1 minute ago, jagsba said:

It just happened. And he's been harrassing to get it to 27.

I mostly saw an opportunity to bring the ace definition back 😂

19 minutes ago, jagsba said:

It just happened. And he's been harassing to get it to 27.

I'll get the other half in when I get done finally doing my taxes.

51 minutes ago, Micanthropyre said:

I'll get the other half in when I get done finally doing my taxes.

Arc dodging Dooku is weird. Plan for the decloak and if he stays cloaked, don’t shoot him.

Arc dodged.

5 hours ago, jagsba said:

It just happened. And he's been harassing to get it to 27.

are we Facebook friends?

13 minutes ago, catachanninja said:

are we Facebook friends?

...yes?

And now meeting for the first time... this hour....

34 minutes ago, jagsba said:

...yes?

Why can't i remember?!

17 minutes ago, catachanninja said:

Why can't i remember?!

You also gave me a spectrum cruiser at worlds

I have 327 lists for wave 3 hyperspace trials so far.

One problem that is by now very obvious is the worse signal/noise ratio. 7 factions mean that nothing is really frequent anymore. Personally I think that's great, but it spreads the meta wide open and makes preparation much harder.

3 hours ago, GreenDragoon said:

I have 327 lists for wave 3 hyperspace trials so far.

One problem that is by now very obvious is the worse signal/noise ratio. 7 factions mean that nothing is really frequent anymore. Personally I think that's great, but it spreads the meta wide open and makes preparation much harder.

Just means you need more data before things start to filter through. We've got a lot of Trials left to go.

Also... I wouldn’t really mind if the meta actually is wide open. Seems like the preferable game state to me.

I forgot who made the point recently-ish about misleading data. On a totally unrelated note, does anybody know who's in charge of the pinksquadron website?

I'm asking because the top10 hitlist is a strange mix of data and undisclosed opinion. Of course the reason I started looking in the first place was the fact that he mentions 5 Awings. I have now 12 hyperspace trials for wave 3, 472 lists and 76 in cuts (16% conversion), not counting milwaukee yet where we'll possibly get another 59 lists.

The hitlist does not e.g. mention 2xT70 and A-wing. But my analysis shows that 3 of the 6 showings made cut, and that's pretty good. ARC+2xDelta is btw one of the best, with 4/10 making cut, and one of the most frequent lists.
This time I count 177 different ship combinations. Half of them were played just once, 50 had 2-3 showings, and the remaining 39 had 4 or more. Those 39 add up to 239 lists (56%). So, half of what's played is some stuff, and half is at least to some degree known. 45 of the 76 (59%) lists in the cut are coming from that more solved/shared part, which is just above the 56%.

The cut had 51 different combinations of ships. 45 of these 51 different list in the cut had a conversion rate above 16%, and 39 lists had more than 20%. These 39 add up to 62/76 lists that made cut (82%), and these lists had on average a conversion rate of 42%.

If I filter out the lists which were played just once, I get 36 lists in cut out of 97 played total, and filtering for those played more than twice, the ratio stays very similar at 22/61 (36% conversion). This last filtering brings the 51 lists from cut down to just 6!

1. 6 TIE fighters (5/19 cut)
2. 2 Aethersprites + 1 Arc (4/10)
3. XYYU (4/13)
4. YT1300+Y (3/8)
5. 2 T70+RZ2A (3/6)
6. XXXU (3/5)

Those 6 lists have a somewhat consistent conversion rate of 36%.

Much worse is for example the 5 Awings list with 1 making cut out of 6 showings, which is where I got interested in the first place. It's still worse than eg the 3 T70+Awing list (1/3) or 4awings+T70 (1/3) or 3 aethersprites (1/5)

(E: SECOND) best imperial is a salad (Advanced, /LN, reaper, striker) with 2/8. Best FO is /vn+/sf+/fo at 2/4 followed by /fo + 2/sf + upsilon at 2/6.
Second best Republic is 1 aethersprite + 3 arcs (2/8). Best separatist is tricky to say: b22+infiltrator+2vultures had 1/5, b22+2 infiltrators had 1/2. Low numbers, so doesn't mean anything really.

Low numbers, but that TIE Swarm is looking very good @GreenDragoon. Get ready for TIE Swarm mirrors everywhere, because the people are paying attention. There were much more TIE swarm players at the event I went to than I was expecting. It's going to drag the numbers back down a bit, as more people play and flop, but it's a solid performer.

I'm thinking though the best CIS is going to be Dooku, Maul, Grievous.... at least short term. It's surprisingly resilient.

2 minutes ago, LagJanson said:

Low numbers, but that TIE Swarm is looking very good @GreenDragoon. Get ready for TIE Swarm mirrors everywhere, because the people are paying attention. There were much more TIE swarm players at the event I went to than I was expecting. It's going to drag the numbers back down a bit, as more people play and flop, but it's a solid performer.

I'm thinking though the best CIS is going to be Dooku, Maul, Grievous.... at least short term. It's surprisingly resilient.

The 'best imperials' is if course wrong, the salad is the SECOND best as the whole list is lead by the tie swarm!

1 hour ago, GreenDragoon said:

• 6 TIE fighters (5/19 cut)
• 2 Aethersprites + 1 Arc (4/10)
• XYYU (4/13)
• YT1300+Y (3/8)
• 2 T70+RZ2A (3/6)
• XXXU (3/5)

So, 3/6 Rebel, 1 Imperial, 1 Republic, 1 Resistance? Not too shabby, for game balance (4 factions in top 6 lists).

Rebels Strong, Bad Guys weak.

2 minutes ago, svelok said:

So, 3/6 Rebel, 1 Imperial, 1 Republic, 1 Resistance? Not too shabby, for game balance (4 factions in top 6 lists).

Rebels Strong, Bad Guys weak.

Is conversion rate a good or even 'the' metric for list performance? I think it's better than just looking at the winning list or for example only top4. But ideally it should be cross referenced with win-loss ratio. And of course consistent outliers are also informative.

1 hour ago, GreenDragoon said:

1. 6 TIE fighters (5/19 cut)
2. 2 Aethersprites + 1 Arc (4/10)
3. XYYU (4/13)
4. YT1300+Y (3/8)
5. 2 T70+RZ2A (3/6)
6. XXXU (3/5)

6 minutes ago, svelok said:

So, 3/6 Rebel, 1 Imperial, 1 Republic, 1 Resistance? Not too shabby, for game balance (4 factions in top 6 lists). Rebels Strong, Bad Guys weak.

...and, for the moment, "(this ship) is OP because I felt bad losing" yells have fallen away... bonus!

1 minute ago, GreenDragoon said:

Is conversion rate a good or even 'the' metric for list performance? I think it's better than just looking at the winning list or for example only top4. But ideally it should be cross referenced with win-loss ratio. And of course consistent outliers are also informative.

It's a metric. I think it's good one, but doesn't tell the complete story. There are a lot of variables but two really big ones are:

• Player skill
• match ups during swiss (incl number of mirrors)

I agree it's better than just counting the number of copies make top cut.

1 hour ago, LagJanson said:

Low numbers, but that TIE Swarm is looking very good @GreenDragoon. Get ready for TIE Swarm mirrors everywhere, because the people are paying attention. There were much more TIE swarm players at the event I went to than I was expecting. It's going to drag the numbers back down a bit, as more people play and flop, but it's a solid performer.

I'm thinking though the best CIS is going to be Dooku, Maul, Grievous.... at least short term. It's surprisingly resilient.

I expect Dooku/Maul to be more of a thing for people who pay attention to the Games and Stuff Trial. Both of them made top 16 and one made the final. It does have a rough matchup against the TIE Swarm, but sometimes it pops 3 ties a round (like what happened in the Top 16).

Upon a day of reflection, I don't think scum is bad right now, just requires a higher degree of execution. In this meta, I would still say that Boba/Emon is the best (or 2nd best) two ship list. Bombs plus reinforce is pretty powerful. I missed the cut by a total of 3 HP spread across 3 games and I played poorly for 2 of those. If you want to run scum, you are probably still "wrong" because you give yourself more opportunities for mistakes and Rebels really don't require much thinking. But I went in knowing I was bringing something hard to play, so keeping the 4-2 memes alive was great.

3 minutes ago, LagJanson said:

It's a metric. I think it's good one, but doesn't tell the complete story. There are a lot of variables but two really big ones are:

• Player skill
• match ups during swiss (incl number of mirrors)

I agree it's better than just counting the number of copies make top cut.

I think that looking at what % of lists go 4-2 or higher may be a better metric than simple conversion (or maybe setting a MOV threshold for 4-2). It seems like you can get a lot of variance with how close MOV can be now that the difference between the cut and 6 places out can only be 80 points.

9 minutes ago, LagJanson said:

It's a metric. I think it's good one, but doesn't tell the complete story. There are a lot of variables but two really big ones are:

• Player skill
• match ups during swiss (incl number of mirrors)

I agree it's better than just counting the number of copies make top cut.

Which would involve some sort of world wide ranking system.

Which would require list fortress to import that finality functionality

Both of which require additional data/effort.

Is there a better metric using the available dates, is a more interesting question to me.