The Jinteki Replicating Perfection Project, Pt2: Agenda
This article was intended to be part of a series that never really progressed beyond the concept stage, so don’t go looking for Pt1 at this time. I may come back to the deck and the series in the future though the former is more likely than the latter. Still, I thought that the notion of Action Investment in scoring various Agenda schemes made for another interesting facet in considering the Agenda you’re building into a deck, so I’ve decided to post this article anyway.
Calculating the cost to win using different Agenda schemes in Jinteki
There are two schools of thought regarding the best Agenda structure for a Jinteki deck. The thinking goes something like this. If all of the Agenda in a deck (let’s assume a standard 49 card deck here for simplicity) are worth two agenda points, then to win the Runner will need to steal four Agenda cards, maximising the amount of work that they need to do compared to a deck including three-point Agendas, where a lucky set of runs could net a win in only three successful runs. Let’s call the first scheme the “All 2s” and the second the “3/2/1 split”, or just “split” for short.
Note that neither scheme is about the reduction of the number of Agenda cards in the deck. To achieve the requisite 20-21 points of Agenda in a 45-49 card deck, the All 2s scheme needs to include ten Agenda cards, thus necessitating the importation of a Neutral two-point Agenda. To achieve the same number of points using a split scheme, the deck must contain either 2× 3-pt, 6× 2-pt and 2× 1-pt (Split1), or 3× 3-pt, 4× 2-pt and 3× 1-pt (Split2) Agendas, in either case totalling ten cards as well.
Extending this slightly, I will also consider one more Split scheme: an extension of Split2 into 11 Agenda cards - where one might run 2× False Lead and 2× Gila Hands Arcology, say - to be 3× 3-pt, 4× 2-pt, 4× 1-pt (Split3).
Other schemes, such as a 3/3/1 split employing only Neutral three-point Agenda could comprise 6× 3-pt and 2× 1-pt Agendas, reducing the total number of Agenda cards to only eight. I’ll ignore this scheme in the analysis. There’s an additional consideration available since Opening Moves (which post-dates this article) in using Clone Retirement as the 1-pt Agenda, but this further complicates the calculations
The All 2s scheme is then predicated on making the Runner reliant upon four successful attack runs - barring additive effects like R&D Interface or Medium which allow more than one card to be accessed at a time. We can ignore the additive access effects for the purpose of the analysis due to the randomness of the order of cards in the deck, and the fact that the number of cards in the first two schemes is exactly the same.
One consideration in choosing between the two schemes was put forward by Geoff Hollis on boardgamegeek.com, suggesting that the All 2s should be favoured by Jinteki Personal Evolution (PE) decks, as it maximises the number of Agenda needing to be stolen by the Runner and so increases the amount of Net damage (Jinteki’s raison d'être) that they will take in the course of the game and so increase the chances of a flatline victory. The argument then ran that Fetal AI had little place in a Jinteki Replicating Perfection (RP) deck, and that is an assumption that I am going to take on board as well.
Jinteki RP is not about flatlining the Runner. Setting aside the controversial statement just made, I think that RP is about constraining the number of runs that the runner can make in the course of the game, and so is more intent on reducing the reward that the Runner receives for the runs that they do make. To adopt Hollis’ later notion of work compression, where PE applies work compression against the Runner’s cards in hand, RP applies work compression directly against the Runner’s clicks per turn. PE seeks to force the Runner into a situation where the work compression against their cards in hand forces them to make a decision between risking a flatline by running with less cards than they can potentially lose (running unsafely to potentially score an Agenda that would otherwise win the Corp the game), or permit the Corp to score an potential Agenda. RP instead seeks to engineer a situation where the Runner simply does not have enough clicks within which to make the crucial run required to score the potential Agenda.
Neither Identity’s strategy necessarily encompasses fast advance through cards like San San City Grid, Biotic Labor or (natively) Trick of Light. Both, very likely, seek to rush early Agenda to increase the work compression on the Runner. Hollis’ observation is that as the Jinteki Corp approaches match-point, compression increases on the Runner as they can no longer afford to allow a potential Agenda to slide, thus giving the game to the Corp.
Keeping in mind that the argument then between the two schools is that one scheme maximises the amount of work that the Runner needs to do, the question in my mind is how much work does the Corporation need to do to win the game by scoring Agenda? What is the action investment (ai) required to score a winning Agenda?
Comparing the All 2s against the Split scheme
Jinteki’s (native) Agenda are Braintrust (3/2), Nisei Mk II (4/2) and Fetal AI (5/2). Additionally, to meet the requisite total of 20 agenda points in an All 2s scheme, one must include either a Private Security Force or Corporate War Agenda, both of which are (4/2). Utilising the Split scheme requires the importation of either Priority Requisition or Executive Retreat, both being (5/3)s. as well as, say, False Lead (3/1).
This makes calculating the action investment for the All 2s scheme moderately easy. The deck composition would then be: 3× (3/2), 4× (4/2) and 3× (5/2). Averaging the advancement costs of the scheme results in an average of four advancements per Agenda scored, and if we choose to count the installation (action) of the Agenda card as well, then we have four installs to get the four required Agenda, plus 4× 4 advancements (adv) to total 20 ai to win by Agenda.
How does this compare against the Split scheme? This is more complex to evaluate, as not only do we have different combinations of Agenda leading to a winning total, but a variable number of advancements required depending on the combination, and also a variable number of cards which could comprise a winning score. Ignoring Fetal AI as an option and choosing only Braintrust and Nisei Mk II as the Agenda in the scheme means that a two-point Agenda will cost, on average, 3.5 actions to score. This gives the following ai values for the split schemes:
for a 3/3/3 win: 3 i, plus 3× 5 adv
= 18 ai (1)
for a 3/3/2 win: 3 i, plus 2× 5 + 1× 3.5 adv
= 16.5 ai (2)
for a 3/3/1 win: 3 i, plus 2× 5 + 1× 3 adv
= 16 ai (3)
for a 3/2/2 win: 3 i, plus 1× 5 + 2× 3.5 adv
= 15 ai (4)
for a 3/2/1/1 win: 4 i, plus 1× 5 + 1× 3.5 + 2× 3 adv
= 18.5 ai (5)
for a 3/1/1/1/1 win: 5 i, plus 1× 5 + 4× 3 adv
= 22 ai (6)
for a 2/2/2/2 win: 4 i, plus 4× 3.5 adv
= 18 ai (7)
for a 2/2/2/1 win: 4 i, plus 3× 3.5 + 1× 3 adv
= 17.5 ai (8)
for a 2/2/1/1/1 win: 5 i, plus 2× 3.5 + 3× 3 adv
= 21 ai (9)
In Split1, (1), (6), & (9) are not possible and can be discarded. Split2 cannot result in (6), and Split3 can see any of the above results as a possibility. Assuming that all of the above winning outcomes were equally likely then we could calculate the average ai for the Split1 scheme as (16.5 + 16 + 15 + 18.5 + 18 + 17.5) / 6 = 16.9167 ai, Split2 results in (18 + 16.5 + 16 + 15 + 18.5 + 18 + 17.5 + 21) / 8 = 17.5625 and Split3 has (18 + 16.5 + 16 + 15 + 18.5 + 22 + 18 + 17.5 + 21) / 9 = 18.0556 ai. All split schemes seem then, on surface, to be more efficient for the Corp to win by Agenda than the All 2s’ 20 ai, but each becomes progressively heavier in ai.
Of course the different outcomes are not equally likely, as the distribution of 3-, 2-, and 1-pt Agenda in the schemes is not uniform. So we need to modify the ai per outcome (Oai) by the likelihood of that outcome emerging; I think a simple and valid way is to weight the ai per outcome based on the number of ways it can occur, divided by the total possible outcomes. I.e. (nOai) / m
where n is the number of ways that the winning outcome is possible, namely the sum of combinations of composite Agenda, and m is the total number of combinations of all winning outcomes. This gives us n values of:
C22 + C61 = 2! / (2-2)! 2! + 6! / (6-1)! 1! = 7 (2’) C22 + C21 = 3 (3’) C21 + C62 = 17 (4’) C21 + C61 + C22 = 9 (5’) C64 = 15 (7’) C63 + C21 = 22 (8’)
in Split1 (73 total possible winning outcomes) and:
C33
= 1
(1’’)
C32 + C41
= 7
(2’’)
C32 + C31
= 6
(3’’)
C31 + C42
= 9
(4’’)
C31 + C41 + C32
= 10
(5’’)
C44
= 1
(7’’)
C43 + C31
= 7
(8’’)
C42 + C33
= 7
(9’’)
in Split2 (48 total possible winning outcomes). Split3 has 60 total possible winning outcomes:
C33
= 1
(1’’’)
C32 + C41
= 7
(2’’’)
C32 + C41
= 7
(3’’’)
C31 + C42
= 9
(4’’’)
C31 + C41 + C42
= 13
(5’’’)
C31 + C44
= 4
(6’’’)
C44
= 1
(7’’’)
C43 + C41
= 8
(8’’’)
C42 + C43
= 10
(9’’’)
This finally results in:
(7×16.5 + 3×16 + 17×15 + 9×18.5 + 15×18 + 22×17.5) / 73 ~ 16.98 weighted ai (aiw) for Split1
(1×18 + 7×16.5 + 6×16 + 9×15 + 10×18.5 + 1×18 + 7×17.5 + 7×21) / 48 ~ 17.44 aiw for Split2
(1×18 + 7×16.5 + 7×16 + 9×15 + 13×18.5 + 4×22 + 1×18 + 8×17.5 + 10×22) / 60 ~ 17.95 aiw for Split3.
Given that my JRPP deck is intending to win by scoring Agenda rather than bouncing the PE Identity’s Net damage off the Runner’s head, I want to choose the least action intensive scheme for my Agenda structure. This suggests that Split1 is the best option.
Note that this is an idealised analysis in which I assume that I’m not going to need to sacrifice agenda points to further my victory. In other words, these values are based on not ever forfeiting False Lead to force the Runner to lose actions. The analysis could be extended to accommodate this into the calculations (determining weighted ai for winning outcomes based on forfeiting one, two or three False Leads) but this introduces a further complexity that I am going to ignore. Instead, let’s consider that for the purposes of the analysis here that any scored and forfeited False Lead is instead actions invested towards making the Runner lose actions in a subsequent turn.
As an adjunct to this consideration, then, False Lead might be thought of as a particularly bad investment. Counting 1 i + 3 adv = 4 ai to force a loss of only 2 actions on the Runner is a net loss of two relative actions for the Corp, especially if actions invested were to translate directly into progress towards victory. What makes False Lead a worthwhile investment, after all, is the timing of the forfeit to make the Runner lose actions at a critical juncture in the game (such as when they might otherwise steal a three point Agenda which the Corp can score next turn). This then translates into an indirect contribution to the victory.
Further analyses present themselves based on this notion. One consideration might be to try and determine the average number of successful runs that the Runner would need to make to steal enough agenda points to win. Another would be to take the aforementioned Clone Retirement Agenda from Second Thoughts into consideration. This might alter the ai for scoring a 1 point Agenda, by reducing it to 2.5 assuming that an equal number of (2/1) and (3/1) Agenda are included in the scheme, or down to just 2 if only Clone Retirement is included. Exercise for the reader, perhaps?
One last scheme presents itself from Second Thoughts onward; a 2/1/1/1/1/1 structure comprising Fetal AI (or Braintrust), False Lead, Gila Hands Arcology, Clone Retirement, and Profiteering (also from Second Thoughts) along with Unorthodox Predictions (True Colors?) Currently this can only be affected by including at least one Braintrust with the three Fetal AI (or v.v.) thus: 2/2/1/1/1/1, but Unorthodox Predictions will allow the full spread of 3×2-pts & 14×1-pts to meet the requirements of a 45-49 card deck. Clearly, however, such a scheme is better suited to Jinteki Personal Evolution than Replicating Perfection.
Post-script: While the initial design of the Jinteki Replicating Perfection Project deck did indeed initially adopt the most efficient scheme, Split1, it evolved through testing and attempting to refine the design to eventually swap from a 2×3-pt/6×2-pt/2×1-pt structure to a slight variant that dropped one Braintrust for two Gila Hands Arcology, increasing the Agenda count by one, while striving to fit in additional economy into a very tight design. The deck has been shelved for some time now, but sustained a better than 50% win rate over time, despite a very negative patch resulting from an Ice deficiency in one version, and weak credits in a subsequent rebuild. I’m planning to return to this very appealing identity and the JRPProject in the near future, though.