Icebreaker analysis circa Future Proof
Please have a look at the first essay in this series for some of the technical details.
Results are “current” up to Future Proof. Notable points here are the inclusion of the following new pieces of Ice (given that I included A Study in Static partially - based on existing spoilers at the time - in the last analysis):
- Tyrant (Str 4 Barrier (0 subroutines) - ASiS);
- Uroboros (Str 4 Sentry - ASiS);
- Data Hound (Str 2 Sentry - HS);
- Salvage (Str 0 Code Gate - HS);
- Whirlpool (Str 1 Trap - HS);
- Burke Bugs (Str 0 Sentry - FP);
- Eli 1.0 (Str 4 Barrier - FP); and
- Flare (Str 6 Sentry - FP).
One of the key things here is that most subtypes of Ice received a new low to very low strength piece of Ice. I expect that this will impact on the average costs to break and drop the overall numbers. The addition of Tyrant will impact on the modelling in that it costs all Fracter Icebreakers 0 credits to pass, as is the case for Woodcutter (included in the previous modelling) and Salvage. This is because I am only modelling them as they come first rezzed, without advancement counters. I’m uncertain what would be a decent approach to modelling them appropriately, as with only one subroutine for an advancement counter, they could be moderately easy to break (except perhaps Peacock vs Salvage with an odd number) but with additional subroutines added they can become quite expensive. The ‘quirk’ that Tyrant and other 0-subroutine Ice throw into the analysis is that they do not contribute to the average cost to break for that set of Ice that they belong to, which of course lowers the overall numbers somewhat artificially.
Additionally, and perhaps more importantly, the following new Icebreakers have been added to the mix:
- Force of Nature (Str 1 Decoder - ASiS);
- Creeper (Str 2 Killer - HS);
- Faerie (Str 1 Killer - FP); and
- Darwin (Str X AI - FP).
A key highlight here is that subsequent to Core there has been absolutely no new Fracter added to the mix, and so we’re now using very old tech to batter down the Barriers. Hardly very future proof, Runners! Mind you, Corroder being the go-to Barrier breaker both in and out of the Anarch faction means that for the most part, there hasn’t been much need for additional development in this part of the cyberwar.
One major correction to the previous analysis is that I believe that I incorrectly modelled Creeper as 1 credit to break a sentry subroutine, instead of the far worse 2 credits.
Additionally, I am not modelling Deus X in the system I’m using as I don’t currently model the (sub-)subtypes of Ice, such as AP, and so it doesn’t work well with my program. I do intend to extend the functionality down the track to handle more subtypes, but Deus X still doesn’t make a good fit into the analysis as its cost to break is a trash of the program rather than simply spending credits.
Darwin poses a problem to the analysis due to its variable strength, and so I decided to go with a simplifying assumption that X = 3. This perhaps represents an ideal situation for the Runner of being able to spend 1 at the start of the turn to add a virus counter to Darwin, then play Surge as an action preceding the run, lifting the Icebreaker to Str 3 (and then assuming this remains the case for subsequent analysis at higher strength due to The Personal Touch, Dinosaurus, or The Helpful AI. Note that the cost of placing the virus counters on Darwin is not modelled in the costs to raise or break, which I think is fair given that Surge is a very likely option in a deck that wants to run this AI Virus.
One further thing to note from the last analysis to now is that after publishing the essay, despite the long-odds of being able to draw into and play two copies of The Personal Touch on any one Icebreaker, things have changed substantially. The release of Cyber Exodus gave us the aforementioned Dinosaurus, which spawned deck archetypes such as Yog.0saurus, Femmesaurus, etc. all banking on the +2 Str from the console to lift the base strength of the Icebreaker du jour. Replicator decks (no matter how effectiive overall) have also made the possibility of double Personal Touching an Icebreaker of choice a real option in the course of a game.
I’m going to retain the original breakdowns in the following case studies, listing them first but in a darker font. There’s a couple of reasons I want to do this, firstly to be able to show the order of the breakers as they were circa Cyber Exodus/ASiS and also to be able to
Fracters vs Barriers
First let’s look at the Fracter breakers vs Barrier Ice:
> sortBy (compare `on` (snd)) $ zip barrierBreakers (mapAcross2 (map averagecost barrierBreakers) barrierIce)
[("Battering Ram",3.857142857142857),
("Corroder",3.857142857142857),
("Snowball",4.571428571428571),
("Aurora",5.714285714285714),
("Crypsis",6.428571428571429),
("Morning Star",29.285714285714285)]
[("Battering Ram",3.3333333333333335),
("Corroder",3.4444444444444446),
("Snowball",4.111111111111111),
("Aurora",5.111111111111111),
("Crypsis",5.777777777777778),
("Morning Star",22.88888888888889),
("Darwin",56.22222222222222)]
There is no movement at all in the orders, but the slightly lower overall average strength of the Barriers faced means that the numbers have come down across the board. Darwin at the assumed strength of 3 is worst of the lot due to being ineffectual against a number of Barriers.
> sortBy (compare `on` (snd)) $ zip barrierBreakers (mapAcross2 (map averagecost barrierBreakers1) barrierIce)
[("Corroder",3.142857142857143),
("Battering Ram",3.2857142857142856),
("Snowball",3.857142857142857),
("Aurora",5.142857142857143),
("Crypsis",5.571428571428571),
("Morning Star",15.142857142857142)]
[("Battering Ram",2.7777777777777777),
("Corroder",2.7777777777777777),
("Snowball",3.4444444444444446),
("Aurora",4.666666666666667),
("Crypsis",5.0),
("Morning Star",11.88888888888889),
("Darwin",45.55555555555556)]
Where Corroder snuck ahead on average here, its now a tie for best Fracter between Battering Ram and Corroder, with the former being listed first purely on alphabetic criteria. Snowball loses some ground here, being 0.57 credits behind Battering Ram before, and now 0.67 credits off the mark now. Aurora remains execrable although still a better bet than Crypsis, and Morning Star’s weakness against big Barriers still leaves it lagging behind all but Darwin.
> sortBy (compare `on` (snd)) $ zip barrierBreakers (mapAcross2 (map averagecost barrierBreakers2) barrierIce)
[("Morning Star",1.0),
("Corroder",2.5714285714285716),
("Battering Ram",2.7142857142857144),
("Snowball",3.142857142857143),
("Aurora",4.571428571428571),
("Crypsis",4.857142857142857)]
[("Morning Star",0.8888888888888888),
("Corroder",2.2222222222222223),
("Battering Ram",2.3333333333333335),
("Snowball",2.7777777777777777),
("Aurora",4.222222222222222),
("Crypsis",4.333333333333333),
(“Darwin”,24.0)]
As expected Morning Star++ again moves to the front against unenhanced Barriers, although again the cautionary word against facechecking an advanced Hadrian’s Wall. Of course, nobody advances Ice now do they? Likewise I guess it is fair and valid to point out that it is more likely that Morning Star++ is in fact a normal Morning Star being supported by one or more Datasuckers, and as such the Hadrian’s Wall+ is only going to cost one more virus counter to break through. The odd average value for Morning Star (which costs 1 to break any number of subroutines, so should be 1.0 average against every Barrier in the analysis) is an artefact of the previously mentioned quirk that Tyrant throws into the analysis. As mentioned previously, I haven’t modelled Tyrant and the other 0-subroutine Ice with any advancement counters at the moment. Darwin at strength 5 still fails against Heimdall 1.0 and Hadrian’s Wall, so rates pretty poorly here.
There are only nine Barriers in the environment, circa Future Proof. At present the average rez cost of Barriers has jumped marginally from ≈4.86 to ≈4.89 credits, so almost no movement there. Likewise the addition of Tyrant and Eli have had little impact on the average strength of ≈3.89 (previously ≈3.86). So the indicators here are that Barriers are approaching the O:NR ‘template’ of costing one more credit to rez than strength of the Ice.
Decoders vs Code Gates
Decoders get a new member to the set in the Force of Nature Anarch Icebreaker, as well as the Darwin AI Breaker.
> sortBy (compare `on` (snd)) $ zip codegateBreakers (mapAcross2 (map averagecost codegateBreakers) codegateIce)
[("ZU.13 Key Master",4.666666666666667),
("Peacock",5.0),
("Gordian Blade",5.777777777777778),
("Crypsis",6.555555555555555),
("Yog.0",66.66666666666667)]
[("ZU.13 Key Master",4.0),
("Peacock",4.416666666666667),
("Force of Nature",4.666666666666667),
("Gordian Blade",4.916666666666667),
("Crypsis",5.666666666666667),
("Yog.0",58.333333333333336),
("Darwin",59.333333333333336)]
ZU.13 maintains the top spot of unenhanced Decoders, but perhaps surprisingly Force of Nature edges out the Gordian Blade as third best Decoder. Darwin at only 3 strength isn’t much up to snuff against the original suite of Code Gate Ice, and the addition of Salvage has made little impact on the relative order of the original Decoders.
> sortBy (compare `on` (snd)) $ zip codegateBreakers (mapAcross2 (map averagecost codegateBreakers1) codegateIce)
[("ZU.13 Key Master",3.7777777777777777),
("Gordian Blade",4.222222222222222),
("Peacock",4.777777777777778),
("Crypsis",5.666666666666667),
("Yog.0",33.333333333333336)]
[("ZU.13 Key Master",3.25),
("Gordian Blade",3.5833333333333335),
("Force of Nature",3.9166666666666665),
("Peacock",4.25),
("Crypsis",4.916666666666667),
("Yog.0",25.0),
("Darwin",27.166666666666668)]
Just one Personal Touch stirs things up a little, with GB leapfrogging FoN, while Peacock is overtaken by both. Key Master is still the master, but the difference narrows even more between the two Shaper Decoders from half a credit to just over a quarter. Darwin remains worst, behind even Yog.0
> sortBy (compare `on` (snd)) $ zip codegateBreakers (mapAcross2 (map averagecost codegateBreakers2) codegateIce)
[("Gordian Blade",2.888888888888889),
("ZU.13 Key Master",3.0),
("Peacock",3.888888888888889),
("Crypsis",4.777777777777778),
("Yog.0",11.444444444444445)]
[("Gordian Blade",2.4166666666666665),
("ZU.13 Key Master",2.5833333333333335),
("Force of Nature",3.25),
("Peacock",3.4166666666666665),
("Crypsis",4.166666666666667),
("Yog.0",8.583333333333334),
("Darwin",11.083333333333334)]
Again with a Helpful AI or Dino support, GB overcomes the ZU.13 by a whisker, finally offsetting the higher raise cost against several mid-high strength Code Gates. FoN holds third, while Crypsis fails to close the gap on Peacock, Yog.0 lags but still runs better than Darwin.
The average rez cost for Code Gates remains lowest of all the classes of Ice; where it was previously at ≈3.67 credits this has now dropped to ≈3.167. Code gates have moved from an average strength of ≈3.78 circa Cyber Exodus to a new lower value of ≈3.167 also. So pound for pound you get the strength you pay for in credits, with Code gates, which is a disappointing development for this class of Ice.
An interesting experiment came to me here, what happens if Cell Portal is removed from the mix of Ice, being such a situational card that not even Jinteki decks run it particularly regularly. I’m just going to report Decodersaurus’ in this case, as the most interesting result:
[("Yog.0",0.2727272727272727),
("Gordian Blade",2.0),
("ZU.13 Key Master",2.3636363636363638),
("Force of Nature",3.0),
("Darwin",3.0),
("Peacock",3.3636363636363638),
("Crypsis",3.909090909090909)]
Yog.0 moves from the back of the pack to the front, blitzing both Shaper Decoders, and interesting also, Darwin becomes a potential option over a Peacock with +2 strength as well, potentially offering even more efficiency with more turns and thus virus counters(until the Corp purges and sets you back to X=0). Notably here though, Force of Nature is a better bet, as you have the flexibility to raise it mid-flight, or employ Datasucker virus counters where only the latter is the option for Darwin.
Killers vs. Sentries
Lastly we come to the Killers and just to re-note that the previously published numbers for Creeper were in error. So I intend to remove these from the original results and treat it as if it were a new entry into the charts instead.
> sortBy (compare `on` (snd)) $ zip sentryBreakers (mapAcross2 (map averagecost sentryBreakers) sentryIce)
[("Ninja",4.357142857142857),
("Femme Fatale",4.928571428571429),
("Creeper",5.142857142857143),
("Crypsis",5.642857142857143),
("Pipeline",6.214285714285714),
("Mimic",43.57142857142857)]
[("Faerie",1.5555555555555556),
("Ninja",4.333333333333333),
("Femme Fatale",4.777777777777778),
("Creeper",4.888888888888889),
("Crypsis",5.555555555555555),
("Pipeline",6.111111111111111),
("Mimic",45.111111111111114),
("Darwin",45.77777777777778)]
Unsurprisingly, Faerie with a 0-to-break subroutine function (at the cost of trashing Faerie at the end of the encounter(!)), is unsurprisingly the most efficient Killer out there; at four credits to completely break an Archer, nothing compares to this. To reiterate a point from last time, Creeper surprises me somewhat by being slightly more effective than Crypsis when “naked”. Ninja is well ahead of the rest of the pack, itself perhaps a surprising thing given that Ninja is a direct reprint of Matador from the Netrunner Classic expansion. Pipeline is appallingly expensive, and when I get around to modelling multiple-Ice run simulations I hope to see something justifying its existence. At this point I think that I am going to stop being redundant and mentioning the lacklustre value of Darwin considered in the '”raw”.
> sortBy (compare `on` (snd)) $ zip sentryBreakers (mapAcross2 (map averagecost sentryBreakers1) sentryIce)
[("Femme Fatale",3.642857142857143),
("Ninja",3.9285714285714284),
("Creeper",4.5),
("Pipeline",4.928571428571429),
("Crypsis",4.928571428571429),
("Mimic",22.5)]
[("Faerie",0.9444444444444444),
("Femme Fatale",3.5555555555555554),
("Ninja",3.8333333333333335),
("Creeper",4.277777777777778),
("Pipeline",4.777777777777778),
("Crypsis",4.833333333333333),
("Mimic",23.27777777777778),
("Darwin",24.333333333333332)]
Disregarding the one-shot of Faerie, Femme Fatale overtakes Ninja as the most efficient (on average) Killer, and Pipeline leapfrogs Crypsis. Creeper maintains the third spot, and Mimic and Darwin bring up the rear, Darwin being marginally worse due to its higher break cost.
> sortBy (compare `on` (snd)) $ zip sentryBreakers (mapAcross2 (map averagecost sentryBreakers2) sentryIce)
[("Femme Fatale",2.7857142857142856),
("Pipeline",3.642857142857143),
("Ninja",3.9285714285714284),
("Creeper",4.071428571428571),
("Crypsis",4.285714285714286),
("Mimic",15.5)]
[("Faerie",0.5),
("Femme Fatale",2.6666666666666665),
("Pipeline",3.5555555555555554),
("Ninja",3.6666666666666665),
("Creeper",3.8333333333333335),
("Crypsis",4.166666666666667),
("Mimic",17.833333333333332),
("Darwin",19.0)]
Here is clear evidence of the potency of the Femmesaurus deck over Dinopipe; with additional Sentry Ice brought into the mix, the difference between the two is 0.89 credits (on average) to break a Sentry. Dinopipe might be a better bet in a Shaper deck where Personal Touches can also appear in faction, whereas in Criminal decks paying Influence points for both Dinosaurus and The Personal Touch (and possibly Personal Workshop) quickly blows out.
Ninja++ remains a solid choice just in touch with Dinopipe, and Creeper is there or thereabouts as well. Crypsis is still a better bet than Mimic, even paying one credit per click to load the virus counter; of course this scales badly to multiple pieces of Ice, and the more credit-rich runner decks become, the more that approximation becomes a tenuous one. In some decks it may make more sense to model the additional cost to load as 1.5 or even 2 credits per click lost to prime the AI Virus.
The average rez cost for sentries drops from ≈4.07 marginally, down to ≈4.056 and the average Sentry strength changed from only ≈3.07 to ≈3.056. One the surface there isn’t a lot of bang for buck in the Sentry subtype, with an average cost of one credit more then strength of Ice (setting aside Archer as a massive counterexample), which is on par with Barriers.
Summation
It remains the case that the prime determinant of the ‘efficiency’ of a breaker in terms of average credit cost to break (the various Ice in its target class) is the raise cost, even with the addition of more types of Ice of lower strength in all classes. The fixed strength breakers remain, as an artefact of my modelling, as the worst option when choosing any of your breaker subclasses. The support that they require lends itself to the program-heavy topology of the Anarch Rig, although Ice Carver also assists, but then it can also assist in any of the above comparisons.
It also bears repeating that the new Weyland Ice is poorly modelled, and while it might seem very weak at this point, may well have its place in an as-yet undiscovered design. Then we might get a better (subjective) measure of the appropriate number of sub-routines with which to approximate their difficulty. This then could impact on the numbers, and would at least pull Morning Star back to a more sensible average of 1.0
Faerie is far and away the best of the Killers, but at the cost of having to retrieve it from the trash or support it with protection cards like Sacrificial Construct. One can hope that it means that abusive combinations that arose from cards like Joan of Arc remain a way off, though if some (reasonable) variant of that card was reprised, it would also lift the power of Crypsis back towards the Bartmoss Memorial Icebreaker is it modelled on. Until then the extra (‘hidden’) cost of Crypsis it ends up being the worst or second worst breaker in every class, while having the flexibility of being the best breaker in all classes (noting that I have not analysed Wyrm at all here and perhaps should in future).
For simple comparison here are the numbers for the Ice again:
Avg(rez cost)
Avg(base strength)
Barrier
≈4.89
≈3.89
Sentry
≈4.056
≈3.055
Code Gate
≈3.167
≈3.167
The second half of the Genesis Cycle didn’t really break too much new ground in terms of either Ice or Icebreakers. Icebreakers have continued the new form of the Cloud breakers of ZU.13 Key Master and Creeper, where having link is an advantage in terms of the MU cost; notably the previously curious phrasing regards costing 0 MU even in hand has been made clear; MU constraints must be obeyed even through the course of installation. As such a Cloud breaker can slip into play even when the MU array is full, as it is 0 MU before the install action occurs, not just after when MU is checked again.
Ice equally has been more variations on existing themes despite continuing the rezzed-only advancable Ice pattern from Woodcutter that has left much of the Weyland player base scratching their collective heads. Something like a Crystal Palace Station Grid might make this Ice a decent option but this would not synergise well with Amazon Industrial Zone, which seems to be universally understood to be part of the solution.
That said, this analysis covers the second arc of a complete expansion cycle. As such expecting revolutionary concepts to emerge between A Study in Static and Future Proof is perhaps asking over much. The new conceptions of Ice and Icebreakers deserves to be more fully explored before breaking new ground.
Have fun with all that!