The Plugged in Tour, Worlds, Andromeda - A Brief Look at Data

Netrunner General
1

The Plugged in Tour was a really great way for us to get tons of information on tournament data in a very brief time span. Although it is unfortunate we don’t have access to the top 8 lists from each event, we can still look at all of the first place lists we have and try and look for any trends.

I will start with runners.

Runner Identities:
Total Lists: 30
Andromeda: 12 ( 40% )
Gabe: 7 (23%)
Kate: 4 (13%)
Exile: 2 (6%)
Chaos Theory: 2
Kit: 1
Noise: 1
Whizzard: 1

Faction Breakdown:
Criminal: 19 (63%)
Shaper: 9 (30%)
Anarch: 2

So no terrible surprises here. The complete domination of Criminal (mainly Andromeda) and the essential nonexistence of Anarch in the post Jackson meta.

Before looking further into this I want to take a little detour and talk about Andromeda. There has been some argument scattered around the various Netrunner forums on whether Andromeda or Gabe is the stronger identity. The Plugged in Tour provides evidence in favor of Andromeda, but let’s look at the other data we have access too.

Worlds was of course won by an Andromeda deck, but in addition 13/16 of the top 16 decks at Worlds were Andromeda. And Andy made up 7/8 of the top 8. Not a single Gabe made it into the top 32!

**EDIT: Since this post went up there now appears to have been one player in the top 32. **

Finally we have the much maligned OCTGN win/loss ratios:

Andromeda (59,67%)
Gabriel Santiago (55,28%)

Basically all of the evidence we have available points towards the superiority of Andromeda, and you will need to have a gameplan against her when building your corp decks. Something to keep in mind for the future is that the OCTGN data had positive predictive power about the dominance of Andromeda throughout the PIT meta.

Specific Cards and Influence

Sure Gamble: 28
Plascrete Carapace: 26
Corroder: 25
Account Siphon: 21
Datasucker: 20
Dirty Laundry: 20
Yog.0: 19
Inside Job : 19
Desperado: 19
R&D Interface: 18
Faerie: 17
Infiltration: 16
Kati Jones: 15
Crypsis: 14
Same Old Thing: 13
Daily Casts: 11
Parasite: 10
Professional Contacts: 9
Ninja: 9
Mimic: 8
Indexing: 7
Stimhack: 6
Atman: 5
Gordian Blade: 3

Meat Damage is a Thing

One thing that immediately jumps out is just how much Scorched Earth impacts the meta. Plascrete Carapace saw ubiquitous play and was clearly a key component of making it to the top table. I would go so far as saying Plascrete is basically a must play for top tier runner decks in the current meta.

The Splashiness of Anarch

Of the top 5 cards that cost influence, 3 of them are Anarch (Corroder, Datasucker, Yog.0). This is especially noteworthy considering that only two Anarch lists are accounted in this. This means that it isn’t a lack of powerful cards keeping Anarch numbers down, but something else entirely. One problem is that the great Anarch cards are very easy to splash, whereas the cream of the criminal crop is expensive (Account Siphon at 4, Desperado and Inside Job at 3).

Corp Identities

Total: 30
Weyland: BaBW: 14 (47%)
HB: Engineering the Future: 7 (23%)
NBN: Making News: 6 (20%)
Weyland: BWBI: 1
NBN: TWIY: 1
Jinteki: Personal Evolution: 1

Poor Jinteki just keeps on trucking as the saddest of the corporations. All of the Core ID’s stood far above any of their newer variations. While Weyland dominated in the PIT, at worlds the top8 was primarily an NBN field (6/8). Others have talked about why this might be in another thread, and I tend to agree with Hollis here:

Ignoring Jinteki, the corps certainly seem to be more balanced, and you can do very well with Weyland, NBN, or HB. Keep in mind that this might change with Restructuring now in the picture.

Specific Cards and Influence
Hedge Fund: 26
Ice Wall: 24
Jackson Howard: 22
Bastion: 19
Enigma : 18
SEA Source: 18
Scorched Earth: 17
Beanstalk Royalties: 17
Wall of Static: 16
Caduceus: 15
Tollbooth: 13
Melange Mining Corp: 13
Green Level Clearance: 13
Snare!: 13
SanSan City Grid: 10
Rototurret: 9
Pop-up Window: 7
PAD Campaign: 3
Neural Katana: 3
Project Junebug: 2

Not all Lists Are Created Equal

One thing to keep in mind when looking at these numbers is that not every PIT winning deck is created equal. The best way to see this effect in action is to note that both Sure Gamble and Hedge Fund were not 30/30 for either side. You might disagree with me, but I consider this to always be a mistake.

Our Savior

Jackson Howard shows his power as the second most played card that costs influence. While Ice Wall was played more, Weyland was also the most played faction, making Howard easily the most splashed card (not that this is terribly surprising). I expect that we will see Jackson’s presence for quite a while.

Wrap Up

Those were just some quick initial impressions. Feel free to throw in your own observations.

Worlds Top32 Lists:

2 Likes
2

Basically all of the evidence we have available points towards the superiority of Andromeda, and you will need to have a gameplan against her when building your corp decks. Something to keep in mind for the future is that the OCTGN data had positive predictive power about the dominance of Andromeda throughout the PIT meta.

I disagree with this-- without the base rates of Andromeda vs. Gabe play, especially among the upper echelons of player skill, we can’t make this conclusion. I would personally expect Andromeda to win substantially more events than Gabe, even though I think Gabe has a higher skill ceiling, as there are simply more people playing Andromeda than Gabe at present.

This might be a local thing, but there is a general consensus among advanced players in my area that Gabe is better than Andromeda at a high level of play. Several of these players use Gabe frequently, and one of them won our second Plugged-In event with him (I took the first one with Kate).

Even under those conditions, Andromeda is still more common than Gabe locally. Something to consider.

3

I maintain that Andromeda is clearly the most consistent runner, in a very strong field of runner archetypes. Any time you anticipate a long tournament (PIT/Nationals/Worlds), I think if you can achieve higher consistency you should be willing to trade a small “raw power” advantage to do so. With exceptional play, Gabe, Exile, Noise, and Kate are all top-tier, but all can be hamstrung by poor draws (for them) or exceptional draws (for the corp). Andy mitigates this.

In the meta spanning nationals -> PIT -> worlds, good to great runners held a huge built-in advantage, so ensuring a more consistent start was a strategically sound move, even if some IDs had “better” built in powers (Gabe, Exile, Noise maybe, Kate maybe). It will be interesting to see if, as corps become less predictable and consequently more powerful, Andy’s consistency advantage fades slightly in favor of IDs with slightly more juice.

I think she will always be a preferred, top-tier ID for big tournament play. You want to maximize the opening as the runner, and there’s no better way to do that than to start with 9 cards.

1 Like
4

Also yes, all of the corps worship at the Temple of Jackson Howard :D.

5

That might be the case, but even if it’s true that doesn’t make him the stronger identity. Unless every tournament you played featured only the top 16 players in the world, all playing out of their skin, the high-level edge doesn’t really make much difference. At Worlds and Gencon the field was huge - even the biggest events necessarily attract bad players and some of them will even make the cut, that’s just variance. You don’t need an elite edge to defeat them, you just need to improve your consistency to avoid those unlucky draws and freak losses to the decks that you should beat every day of the week.

Andy’s starting hand makes a huge impact. You get a flying start and that can be worth more than the “four clicks” her ability is often rated as. The link shouldn’t be underestimated either - it’s effectively worth $1 in every trace, so in the right match-ups has the potential to be more valuable than Gabe’s ability.

Finally, Gabe’s ability is limited to successfully running HQ - so he telegraphs his intentions from before the first card is even drawn. Andy can be more versatile with her strategy and potentially discarding cards in turn 1 can also be an advantage with the amount of cards we are seeing that access the trash.

As the game gets older the numbers will shift even further in favour of Andy, unless HQ access gets a big push. Accessing HQ doesn’t usually win the game for you, accessing R&D and remotes does and Andy can be doing that quicker and more effectively than Gabe.

3 Likes
6

It is not conclusive evidence of course, but it is evidence in favor of Andromeda being stronger. If Gabe was truly stronger we would have predicted him to do better at Worlds especially. The point is that we have no data based evidence for Gabe > Andy, just the personal feelings of some players.

I have heard other folks say this too, and I think it is basically without any merit. The common line of thought seems to be that if you play Gabe perfectly and make no mistakes, then his power is greater. There is a huge gaping problem with this line of thought. If you play Andromeda perfectly then baring bad luck you will win the game, so who cares if his theoretical perfect play power is higher? Additionally I would say that Andy is even more of a sure thing than Gabe is at perfect play, because Andy runs RaD Interface and locks the game down, making lucky draws less of a factor.

2 Likes
7

The fact is, the only way to truly settle the argument is to have the same person play a Gabe deck and an Andromeda deck consistently against the same opposition and have a look at the results. Could do this multiple times with several different people, but the data would be analysed separately to see if there’s a trend. Bulk data like this is just too susceptible to individual player bias, we really have no idea if those running Andromeda would have done better or worse with Gabe. Those people preferred Andromeda initially, for whatever reason, and that’s really the only conclusion; most runners prefer to run Andromeda over Gabe.

My personal thoughts are that they’re awfully close, and it might be down to individual card interactions and matchups to decide the “winner”. The balance could easily shift with a single new card. Certainly the prevailing perception is that Andromeda is the winner, and you have to admit that this does affect who plays with her, especially at tournaments.

8

I don’t agree that this would “settle” it, and it’s not a very scientific approach.
Firstly, it’s not just a case of switching the IDs over, they play in different ways so you have a deck-building element to the test and you’re not really isolating the variable you’re examining. You know this is the case because you say:

Secondly, you can’t conduct each game in a vacuum under the same conditions. With every game played the player learns about his opponent, their choice of cards, the risks to avoid etc. this means that the skill level is not a constant between tests.

OCTGN basically is this experiment, but with multiple players. If you assume that with a sufficiently large data set different players’ skill levels are evenly distributed, the OCTGN data is the largest sample we have and all signs point to Andy being “better”.

Every single proponent of Gabe says things like [quote=“Casanunda, post:7, topic:364”]
My personal thoughts are that they’re awfully close
[/quote]
Or similar comments with no evidential basis.

This is because you don’t understand statistics. YES Andromeda will see more wins because she is played more frequently, what is indicative of power level is what proportion of decks are winning compared to those that are played. If every tournament had 99 Andromedas and 1 Gabe, but Gabe won 25% of tournaments you might be able to say he’s better, since if they were equal you would expect him to win in proportion to how often he’s played (i.e. only 1%).

1 Like
9

This is because you don’t understand statistics. YES Andromeda will see more wins because she is played more frequently, what is indicative of power level is what proportion of decks are winning compared to those that are played. If every tournament had 99 Andromedas and 1 Gabe, but Gabe won 25% of tournaments you might be able to say he’s better, since if they were equal you would expect him to win in proportion to how often he’s played (i.e. only 1%).

Uh, that’s actually what I’m arguing, though I certainly don’t believe that the ratio is as extreme as 99 to 1.

10

I never said that, please read more carefully. I clearly used the word “deck”. Certainly there will be differences in the decks themselves, which certainly could imply that its the deck itself that proves to be better, not the identity. That’s a perfectly valid thing to say (that a typical Andy deck is better than a typical Gabe deck) , I just want to be clear about what we’re saying is better.

I agree, it’s difficult, no question. It will always be very hard to control the opposition, but there are cases where you just have to concede that a particular variable is going to be difficult to control and do your best to control what you can. In this case, controlling the running player seems like the best case scenario, you can at least take runner skill mostly out of the equation (assuming they don’t get significantly better as more games are played). An attempt to play against similar opposition might be the best you can do.

OCTGN is probably the best set to go by at this point, but that’s a pretty tough assumption to sell, to be honest. Larger sample does not mean better, I’m afraid, you simply have too many groups within that set to conclude anything. If I’m to go by OCTGN data, then TWiY is better than Making News. I’m not sure if I’m willing to agree with that.

My lack of evidential basis is roughly the same as your lack of conclusive evidence, I’m afraid. Guess we’ll just have to stick to opinions.

11

Agreed there are a lot of groups and sub-groups within OCTGN, but you can make some educated cuts (as people have done). For instance if you ignore all games involving decks where Influence = 0 as these are probably the core decks out of the box, or at least poorly tuned since we know from prior analyses that there is a correlation between influence and win rate (The Collective being the exception). Games are time stamped so you can track player IDs and apply a post-hoc rating system to rank players and then only consider certain percentiles - that gives you handle on skill level so you could consider only the best players vs. each other.

So, yes there are groups but they’re not impossible to extract or exclude from the sample. Besides, the larger the dataset is, the more reliable it becomes - irrespective of these groups. I get what you’re saying - you can’t tell what’s a good player vs. a bad player, or a tuned T1 deck vs. an experimental draft deck. That’s true, but if the dataset is large enough then skill level mis-matches average out (i.e. there will be as many good vs. bad as there are bad vs. good for any particular ID head-to-head pairing).

Well, there’s no such thing as conclusive evidence. We can only ever speak in terms of statistical confidence, but the evidence for Andy being conducive of a more successful deck is stronger than that for Gabe. The only defence anyone seems to be able to offer Gabe is gut reaction, nothing tangible.

It seems intuitive to me that Gabe would be less successful. The most efficient decks are the ones that maximise the pay-off for doing actions you were going to do anyway (c.f. The Collective). That’s why R&D interface is such a good card - you know the agendas are in there and R&D lock is a thing. Gabe makes the pay-off for hitting HQ slightly better, but hitting HQ doesn’t win you as many games in general.

12

I’ve seen the numbers before, but that was a while ago, and I’m sure the percentages have changed since then. I’d love to see more recent breakdowns, but I’m not sure if anyone has done it; that can be a lot of work. While I know that larger sets are more reliable, and 100000 data points seems like a lot, once it’s separated into the various groups the numbers become much smaller, with a greater degree of uncertainty. All that I’m saying is that since I don’t have the “upper level” breakdown available, I’m hesitant to put much faith into what you could call the “raw” numbers from OCTGN.

I don’t want to come off as being biased, though; I’m just skeptical. I’m willing to concede that Andy is the better identity, I just find the arguments that Andy is better … unconvincing? Suggestive, rather than conclusive? If someone were to provide the kind of breakdown that I suggested before, that would likely be sufficient to convince me, that was really my only point in mentioning my “experiment” in the first place. I’m not trying to be difficult, I’d just like to see something a little more concrete before I jump aboard.

I’ll give Gabe some “theoretical” defense, for discussion’s sake. It’s been discussed many times why things like Datasucker and Desperado are such great cards; they give you “free stuff” for what you want to do anyway (run). I’d even throw John Masanori in there. In that sense, Gabe’s ability seems to put him into that category of great cards. Whether or not you should be running HQ seems to be the sticky point. Then again, 2 of the most powerful crim cards are dependent on running HQ, so it certainly seems like any criminal deck should at least try to get into HQ at some point.

Incidentally, why am I using quotation marks so much? Seems like I’m only capable of “ironically” defending Gabe…

1 Like
13

Exactly. This is not a settled fact or the meeting of some confidence interval, but it is some Bayesian evidence in favor of Andromeda. The probability of Andromeda being the strongest current runner should be updated upwards.

1 Like
14

Evidence that Andromeda is superior to Gabe:

  • Andromeda’s win rate on OCTGN is ~4.5% higher than Gabe’s, with a sample size of over 17000 games of Andromeda and 28000 of Gabe. In more recent data (post opening moves), Andromeda’s win rate is more than 6% higher than Gabe’s (~59% vs ~53%), with a sample size of about 8000 Andromeda games and 5500 Gabe. I will stress that 6% of win rate is a TON of win rate, and the sample size is very large, sufficiently large to bring the measurement error to about 1%.

  • Andromeda had far more plugged in tour wins than Gabe, as this thread described.

  • Andromeda took 7/8 top 8 and 13/16 top 16 at worlds, while Gabe wasn’t in the top 32. That is an amazing level of dominance, even if Andromeda was much more commonly played.

Evidence that Gabe is better:

  • Certain people on the internet said they think its better.
4 Likes
15

To be honest, I care a lot more about the opinion of highly skilled players than I do about OCTGN. I think the respective tournament showings are based on Andromeda being represented much more, not on her being better.

I’d be happy to admit I’m wrong given appropriate evidence-- if, for instance, FFG releases statistics on tournament decks as a whole and Gabe is seeing as much play as Andromeda, then I would be confident that Andromeda is better.

16

Even if Andromeda was played 4 times as much as Gabe, the fact that it hoarded almost all the top spots while Gabe had none is evidence that Andromeda is better.

Also, the fact that all these top players are choosing Andromeda over Gabe in the first place, is evidence that Andromeda is better, because the top players know what they are doing!

Also, the argument that goes ‘Gabe is better, but only at top level play’ is kindof blown out of the water when at the top tournament, Worlds, Andromeda so thoroughly crushes Gabe in performance.

Also, the OCTGN data is clear evidence at a large sample size that Andromeda is better overall, averaged over all skill levels. While this doesnt specifically disprove that Gabe is stronger at high skill, it at the very least means that there is a much higher bar. Gabe has to gain more than 6% in win rate at high skill play in order to catch up with Andromeda.

Maybe if all of the ‘experts’ were claiming that Gabe was better at high skill, you could make a case based on that (though you would be valuing the experience of a handful of individuals over the actual data). But that isnt even the case, since there are probably more experienced netrunner posters who prefer Andromeda.

I think that if someone was initially making a case for Gabe being better at high skill a couple months ago, they should have re-evaluated that by now, as the pluggedi n tour data and then the worlds data have dealt serious blows to that position.

2 Likes
17

I actually care a lot more about the OCTGN data even than my own experiences. People are flawed, and biased, and the clarity of real data of large sample size is just far more compelling. I initially thought NBN:TWIY was worse than MN, but now its clear that they are fairly equal, and if anything its a tiny bit better on average. I can accept when the OCTGN data tells me I’m wrong. The Gabe defenders should probably learn to do the same.

18

I would honestly expect Andromeda to be played more than 4x as much as Gabe. And yeah, Andromeda winning is evidence-- but the fact that evidence exists for something doesn’t mean that that thing is true! We have some pieces of evidence pointing towards Andromeda being stronger, and others pointing towards Gabe being stronger. I believe that the evidence in favor of Gabe is currently more compelling that that in favor of Andromeda, though there’s certainly room for disagreement here.

Again, we don’t even know if that happened. How many people took Gabe to Worlds? Also, while Worlds may be the biggest tournament, it’s far from fully inclusive. For instance I believe literally 0 people from my local meta (which I consider very strong) attended.

I believe all or almost all of the very strong players in my area think Gabe is better than Andromeda. And yeah, I definitely value the experience of a handful of individuals. I have far more faith in the appraisals of elite players than I do in any amount of OCTGN data.

I used to think the same, but it quickly became apparent to me that people in my local meta were generally far-and-away better than the people on OCTGN. I remember being rated #1 in an OCTGN league and attending my first tournament at a new store, only to barely squeak in to top 8 (I think I placed 7th or 8th) out of a fairly small field!

When things like that are the norm it really calls the validity of OCTGN stuff into question.

19

Thats the thing though. None of the evidence points towards Gabe, at all. Your group of pro-Gabe individual anecdotal evidence/opinions counteracts others groups of pro-Andy anectodal evidence/opinions. And then all the data, tournaments, and worlds, point towards Andromeda.

So while no amount of evidence in favor of Andromeda is going to allow you to say she is conclusively, 100% better, the fact that all the evidence is on that side should at least push you strongly toward that side.

Most of the top players play on OCTGN too. If your in-person experience is playing against strong people and your OCTGN experience is playing randoms, then the in person players are going to be way better.
On the other hand, if you specifically went and met with strong players on OCTGN and played them there, then your OCTGN play would be the tougher competition.

Overall, OCTGN is pretty representative of overall Netrunner players.

Let me use the classic example:
Hypothesis: “There is a dragon in my garage!”
“Wow, can I see it?”
“Its an invisible Dragon”. (Explains away visual evidence against the dragon).
“I’ll just listen to it then”
“Its an inaudible Dragon”. (Explains away audible evidence agaisnt the dragon).
“How about I touch it?”
“Its a whatever word means you cant touch it dragon”. (Continues explaining away all evidence).

Hypothesis “Gage is better than Andromeda at high skill”
“How about this OCTGN data of higher Andromeda win rates”
“OCTGN data is bullshit, and OCTGN players all suck” (Explains away OCTGN data).
“How about Andromeda winning more Plugged in Tours”
“Plugged in tour players arent actually that good either” (Explains away Plugged in tour data).
“How about Andromeda taking 13/16 top 16 at Worlds, with no Gabe”
“Everyone played Andromeda over Gabe” (Explains away Worlds data).

At what point does one allow reality to actually impact their opinions here?

Might I assert that the evidence actually supports the opposite conclusion? That actually Gabe is weaker at high level play. That the Worlds data tells us that the top players can successuflly defend against Gabe better than they can against Andy, surviving the opening and reaching lategame scenarios where Andromeda decks are much tougher to deal with than Gabe? I wasnt even going to go that far, but thats the conclusion that the evidence ACTUALLY points at.

1 Like
20

So a problem in your reasoning is that this points towards Andromeda. As far as I can tell the best players think Andromeda is better.

I think doing well at tournaments correlates to player strength. If it does, then clearly the better players think Andromeda is better.