I’m pretty sure that the numbers have been crunched, and there’s not a big difference for rnd drugs between 7 agendas or 12.
GT is a special case because the game is so dominated by if it’s stolen or not (from rnd)
So here’s some half-assed simulation results. This shows the cumulative distribution function for #of accesses required to steal 7 points, assuming you are accessing cards at random from the deck (without replacement). This elides over a number of critical factors, but gives you at least a rough idea of how agenda density and point value interact.
The X axis is the number of accesses required to win, and the Y axis is the probability of such an event occurring. So, the point where a curve crosses p = 0.5 is the median: basically 19-20 accesses for both the “medium” and “heavy” curves vs. 14 for the “light” curve.
The agenda suites were:
Heavy: 3x Vanity Project + 3x GFI
Medium: 6x 3/2, 2x GFI, 1x 4/2
Light: 16x 2/1 or 3/1, 3x 3/2
gt: 1x GT, 3x Vanity Project, 1x GFI
Roughly, the flatter the curve is, the more variability there is. Both the “medium” and “heavy” agenda suites have a median number of accesses of around 19, but “heavy” shows a lot more variability: 10 accesses won against “heavy” in ~15% of trials vs. 7% for medium. On the flip-side, 30 accesses are sufficient to win against 93% of “medium” compositions vs.86% against “heavy.”
Maybe my “light” composition was totally whack but the general intuition of “it takes basically the same number of accesses regardless of composition, but you get more variation with bigger agendas” is about right, with the limit cases of tons of small agendas or very few extra-high-value agendas being somewhat worse.
And just to reiterate the disclaimer: these are totally unrealistic assumptions that aim to just give a visual/intuitive display of the intuitive discussions above, and clearly player skill/strategy will radically shape the number of accesses required to actually win.
4 Likes
Thank you, @internet_potato, for giving me a rough set of statistics to say a shaky ‘told you so.’
So yeah, basically the probability of a Runner blow out win on a lower number of accesses is higher with the big agendas.
That’s on top of the fact that Takeover and Vanity are nearly impossible to just score.
Potato… Can you do the same rough calc with 4x 4/2, 1x 5/3, 3x GFI? This would be the RP spread. Assume no TFP. Also, same but 1x GFI and 3x 5/3s?
Thank you for making the chart, it’s very interesting.
Another way to look at the data might be to look at the number of accesses at which one composition becomes safer than another. For instance, it looks like if the runner gets 10 random accesses, they are more likely to win if it’s the heavy composition than if it’s the light composition; if they get 11 random accesses, the reverse is true. So the eleventh access is when the heavy composition becomes safer than the light one.
It’s also interesting that the light composition is never substantially safer than the medium.
There are 22 points in the light deck.
1 Like
Hard mode engaged
Python script here
agenda suite definitions here
Results in .csv format: Plot these in excel by clicking somewhere in the data and doing Insert → Scatter. It should be smart enough to figure out how to plot it.
If you have python installed, you can make your own agenda suite definitions file and ram them through. It would be interesting to see, for instance, if Weyland with 3 Hostile Takeovers (21 points) is substantially different than with 2 Hostile Takeovers (20 points).
Fixed plot of medium/light/heavy/gt (with 21 points in “light”):
Pretty sure that the reasons “light” looks worse than medium/heavy are:
- Lack of GFI means that there are just more steal-able points
- Presence of 1-pointers means that there are more ways to hit exactly 7 points.
Plot of medium, heavy, and @Orbital_Tangent’s requests (rp = 3 GFI, 1 5/3. rp2=3 5/3, 1 GFI).
This reinforces the “lol GFI is OP wtf” intuition-- look at how much worse the non-GFI (rp2) suite is than the basic RP suite.
2 Likes
Thank you @internet_potato for the data! It’s really interesting to see that the inclusion of GFI is more important than the number of Agendas you include.
With this in mind, it seems the next consideration must be how “score-able” a given agenda is. As we’ve discussed, Vanity Project and Government Takeover (And Mandatory Upgrades FWIW) are very difficult to score, requiring 2+ Runner turns to complete (seven clicks, including install).
Are 2/3s and 2/4s any easier to score than 3/5s, given that without Fast Advance you only need to hold back the runner for one turn to score any of those?
Yeah, in fact, if you have a 21-point, 9-agenda deck (6x 3/2 + 3x 5/3), the median accesses-to-win goes: 16 (no GFI), 17 (1), 18 (2), 20 (3). That’s pretty significant. You can also see that the step from 2->3 GFI is comparatively large (from 18 to 20), probably because you have no way to win on 3 steals.
I thought the only agenda we have to score these days was Breaking News.
In all seriousness, though, I think that protecting an agenda for two turns is very hard to do in the current meta and scoring VP or GT is unlikely in most games (though that may change with Jemison Astronautics and the other Weyland FA tricks coming up in Red Sands).
My opinion: if you’re not fast-advancing, then the advancement cost mainly matters in that it constrains your ability to bluff (modulo 2/1’s and 2/0’s). There was a thread on reddit a little while ago discussing the merits of advance-able traps in the current meta-- most people kind of agreed that it looks suspicious as hell when you IAA something that you know the runner can access, and that even if a runner does hit an advanced trap they can normally play around it to avoid a game-losing mistake (and if they don’t check it, you wasted an entire turn, 2 credits, and a deck slot).
That may have been a bit of a tangent. If we’re going strictly on numbers, then it seems like 3xGFI and then a bunch of 2-pointers and at most one 1-pointer is the safest composition. When you take into account whether you’re going to never-advance everything, whether some agendas benefit from being over-advanced, and whether you have advanceable traps (or cards like Thomas Haas / Sleepers that can bait a run without hurting your board state as much), “difficulty of scoring” gets complicated fast.
The only problem here is that unless you want a lot of 1/3s, you cannot “never advance” enough 2 pointers (unless you are HB or want to use Merger); from the above discussion it would appear that Merger is the worst thing we can add to a deck given the benefits of GFI.
For reference, 3/2s:
- ABT (HB)
- Project Vitruvius (HB)
- Braintrust (Jinteki)
- Philotic Entaglement (Jinteki, Limit 1 per deck)
- Astroscript Pilot Program (NBN, limit 1 per Deck)
- Project Beale (NBN)
- Project Atlas (Weyland)
- Merger (Neutral 1 Influence)
As we can see, only HB has 6x 3/2 available to them (this is assuming we agree that Merger is not worth the influence or increased density it gives the Runner)
I guess, this must just mean that we have to rely on 4/2s to fill in our non-HB glaciers. But since the scoring window is the same as a 5/3 (one runner turn with something partially advanced), does it benefit the corp to slip in a few larger agendas anyway? And as was stated, that’s where the discussion gets complicated.
So, what 4/2s do people like? I’ve noticed a certain love for Corporate Sales Team, and I’m personally fond of Advanced Concept Hopper in my HB decks.
And, do people slip in extra 5/3s? What do you use? I was running Project Wotan in anything with bioroids, and naturally TFP in Jinteki.
I am pretty sure @beyoken did a video on this very topic a few months ago on his Youtube channel (I am at work atm so I cannot link it).
When it came to 5/3s, it boils down to they have to defend themselves otherwise they are not worth the runner only needing to score 3 agendas. That is why The Future Perfects and Global Food Initiatives are the only 5/3s that really get run competitively. The Future Perfect can be overcome with Film Critic but if you do not have that, it is a pseudo 1-pointer for the runner (since every time they access it, they have a 1-in-3 chance of stealing it).
The benefits of 5/3s at this point in time are not strong enough to mitigate the drawback of the runner only needing to steal three agendas to win.
Funny you mentioned me. I have a new video on this topic lined up, planning to release it tomorrow 
2 Likes
I had a phase (you know, like APEX or tequila) where I played one Merger in what I think is the only deck it can be arguably justified - Titan all-in FA so you can avoid the “runner stole one Atlas so sad” issue. Even ran one GFI to “balance” it out. I didn’t believe the numbers.
It is the worst. For every game that extra 3/2 saved me, it lost another two with its downside. Not to mention the play-warping - freaks you out when it is in hand, it gives a massive tell when seen, etc. Above is from a 30 or so game sample of an otherwise solid FA build. Immediately did better w/out it. The negatives are real…
2 Likes
As promised, analysis on common agenda compositions!
5 Likes
I believe a more realistic ‘Light’ is to have 3x Fetal AI, 1x Philotic, and then 12x 2/1 or 3/1, mostly because only Jinteki PE is interested in running it. (For reference, my current testing PE build is 3x of the following: Clone Retirement, House of Knives, Profiteering, Ancestral Imaging. … What, I like watching Runners die to a Komainu.)
For other concerns: I’ve generally seen Kill decks running Explode-A-Palooza over NAPD, because A) Your credit pool matters more than how small theirs is, and besides the credit differential is larger for EAP. B) One less influence.
(Current Boom combo Sync deck spread: Astroscript, 3x EAP, 3x Kitty, 3x BN, 2x Corporate Sales Team. The influence is all used by the kill package, so cannot slot the GFIs.)
The rules are:
- GFI is good. If you have influence available, you should run these.
- 3/2’s are good. You should run all that you can.
- 3/1’s and 5/3’s are both bad, for various reasons. You should only be running these if they give you a benefit that mitigates their problems.
I have broken these rules in three decks. PE breaks rules 1 and 2, and the reason for including 3/1’s is to abuse its ID as much as possible. PE doesn’t ever want to score, so doesn’t care that GFI allows it to win more easily. Jinteki’s 3/2 is Braintrust, which is bad, and we just said that PE doesn’t want to score out that badly, so including it dilutes what you’re trying to do. PE wants the runner to score as many agendas as possible before winning, which 3/1’s happen to be very good for.
Weyland Punitive kill breaks rule 1 because they want to actually deal 3 damage with the Punitive. (This is also the reason for breaking rule 3)
Finally, Weyland Builder of Nations broke rule 1 and 3. (Agendas: 3x Cleaners, 3x Atlas, 3x Hostile Takeover, 1x Oaktown) It didn’t have the influence to include GFIs, and including GFIs actually made the agenda distribution rather odd. It ran 3x The Cleaners, because that card being scored generally determined winning or losing for that deck. (This, by the way, is one of the annoyances about BoN. I might look into rebuilding it again using GFI and no Cleaners, but I suspect it still won’t be good.)
The runner can’t decline to steal Explode-A, mainly.
3 Likes
My basic Weyland suite for a MCC deck is:
3 Project ATLAS
3 Oaktown Renovation
2 Hollywood Renovation
2 Hostile Takeover
Why Hollywood?
Because of Mumbad Construction Company’s limitations in what it can do.
It can only move counters to face-up cards, which Hollywood is. Additionally, Hollywood allows you to put counters onto MCC, so that you can Rez MCC, get counter, IAA Hollywood, putting two more counters on MCC, then move all three counters to Hollywood and score it.
You can score it from hand with MCC if it gets 3 counters.
Install Hollywood, advance advance, move 3 counters. Cost = 8 creds.
You can score it from hand with unrezzed MCC and Dedication ceremony.
End of runner’s turn, rez MCC. First lcick install hollywood, next click dedicate MCC, next click advance hollywood, then move 4 counters. Cost = 14 creds.
Both happen a lot. Also, (rarely), you can do the Hollywood behind Firwall into Project Atlas thing and build a fortress and fast advance and crush a game out instantly.
1 Like