Beyond Brian's Brain

[Apple Bottom]

One way of "taming" an explosive rule is to treat it as a Generations rule and increase the number of generations, until cell death quells cell birth sufficiently to avoid explosion.

Seeds (B2/S) is explosive, of course, as is Brian's Brain (B2/S/G3, aka /2/3). Adding a fourth generation yields a rule (the unnamed /2/4, which I'd suggest may be called "Brain 4") that's less explosive, but still occasionally blows up:

Code: Select all

x = 20, y = 20, rule = /2/4
B.CB.C3.A2.A.B2CA2B$2C2AB3.C.A.2B2.A2.B$3.BA2.2CACA2.2B.AC$AB5.A.A.BC
2.BA3C$.A.BAC.A2.2C2.A4.A$BA4.2BA3.CBA4.A$BA.CB.2BCA2.AB.A$4.C.C.A2.C
B2.2A2BC$CABA4.B.C3.A.AB.B$C.B.3A.AC.B.CB.AB.B$2A4.2A.C.2A.B2A$.B.C2.
BAB.A.3BA.B$.C.C.B2.C.BAB.2AC$2B.A.B.3B.AB2CBAB$B.B.BA3.BAB2AB3C.C$AC
4.C.2BA4.C2A$2BA2.A.2A2.B3.B2CBC$A.B3.3C2A.2ACA2.C$.2AC2.A3.B.B3.C.B$
CA6.B.2AB3.C.BC!

However, adding another generation (ending up with /2/5; "Brain 5") makes the rule sufficiently stable to apgsearch it. It's fairly slow (~0.25 soups/sec for C1), but I managed to complete a haul. The ash is primarily photons.

/2/6 ("Brain 6") is comparable in terms of both stability, search speed and yield.

I've not looked at higher symmetries or tried other search programs, but I'd be curious what others can find in these rules.

[dan_zh]

Thank you very much for this interesting post! I will fiddle around with these rules!

[Rhombic]

b34s345 starts to be stable at or after 7 generations. 345/34/7 seems MUCH more interesting than 345/34/8 but the search is around 200-300 times slower. That is a lot.

[Apple Bottom]

b34s345 starts to be stable at or after 7 generations. 345/34/7 seems MUCH more interesting than 345/34/8 but the search is around 200-300 times slower. That is a lot.

Indeed -- though I think they're both quite interesting. For that matter so's g9b34s345 (345/34/9), which is also quite fast to search.