The Online Life-Like CA Soup Search

[Macbi]

I've just realised that the b3/s23 census on this site has counted more objects in total than Achim Flammenkamp's census (at http://wwwhomes.uni-bielefeld.de/achim/ ... _life.html).

Comparing the frequencies of the 35 most common still lives and oscillators in each:


The anomaly is the ship. Why is it so much more common in the soup search? Maybe one of the reactions producing it takes a lot of space, and so can't happen in the dense toroidal universe Flammenkamp used, but can be formed on the edge of a cloud from the soup search.

[calcyman]

Maybe one of the reactions producing it takes a lot of space

Ships appear almost exclusively in the debris of the B-heptomino, which flourishes in empty space. In toroidal universes, there is a uniform density, and not enough empty space for a B-heptomino to evolve.

[Lewis]

I've noticed when running Day & Night in the Soup Search, occasionally infinite growth patterns (wickstretchers, I think) turn up. These seem to be undetected by the script, ie. the game isn't stopped and the pattern is seemingly grown forever.

Does the Soup Search ever detect these after a while? and if not, will it be fixed in a newer version of the script?

[Nathaniel]

Does the Soup Search ever detect these after a while? and if not, will it be fixed in a newer version of the script?

I believe it does *eventually* stop, though it's been a while since I ran the script on Day & Night. As for the next version of the script, I've just about hit the limit of what I can do with the current script I think -- the next version of the script will likely be based on the WinLife census script, but that's still a while off.

[Lewis]

Does the script detect pseudo-spaceships the same way it does with pseudo-still lifes/oscillators?

Andrzej Okrasinski's census counted around 20 or so 'LWSS on LWSS's: (I'm not sure why they're included though, because the census doesn't take pseudo-still lifes into account):

Code: Select all

x = 5, y = 9, rule = B3/S23
2b2o$2ob2o$4o$b2o2$b2o$4o$2ob2o$2b2o!

Would this be counted as 2 LWSSs or as one new object?

[calcyman]

I seem to recall that the 'LWSS on LWSS' was an error, and didn't actually appear in Okrasinski's searches.

[Nathaniel]

Would this be counted as 2 LWSSs or as one new object?

It would be counted as one (pseudo) object.

[Lewis]

In the results table for standard Life rules, the picture for object 152 (2nd from the bottom) doesn't show up. Is there any reason for this?

Edit: The total soups count for 23/35 has gone to 0, like the Life count mentioned earlier.

[Lewis]

The table on the main page of the soup search has gone strange:

Soup Search.PNG (13.04 KiB) Viewed 18616 times

[Sokwe]

The soup search has found a small period-136/2 c/2 puffer in B35/S23:

Code: Select all

x = 15, y = 18, rule = B35/S23
$7b3o$6bo2bo$9bo$5bo3bo$5b2o3bo$5bo$6bo2bo$6b3o!

Was this puffer already known, and if not, is it the first known infinite growth pattern in this rule?

Here are some spaceships of periods 24, 48, and 112 based on this puffer:

Code: Select all

x = 177, y = 133, rule = B35/S23
12b3o14b3o$16b3o6b3o$16bobo6bobo$16bobo6bobo7$13b2o2bo2bo2bo2bo2b2o$
12bobobobobo2bobobobobo$11bob2obo10bob2obo$11b2o3b2o8b2o3b2o$10bobo2bo
2b3o2b3o2bo2bobo$15bo12bo5$14b2o12b2o$14b2o12b2o$21b2o$21b2o$11b2o3b2o
8b2o3b2o$10b2obobo2bo6bo2bobob2o$11b5o2bo6bo2b5o$6bo4bobo4bo6bo4bobo4b
o$5b3o4b4obo8bob4o4b3o$5bob2o4bob2o4b2o4b2obo4b2obo$5b3o6bo6b2o6bo6b3o
4$21b2o$21b2o$12bo18bo$8b2ob4o14b4ob2o$8bo5bo14bo5bo2$8bo3b2o16b2o3bo$
8b4o9b2o9b4o$8b2o11b2o11b2o2$74bo$73b3o58bo23bo$2bo18b2o18bo30b4o2bo2b
o51b2o23b2o$b2o8b2o8b2o8b2o8b2o30b2o7bo50bo25bo$o2bo5bo3b2o14b2o3bo5bo
2bo33bo3b2o$obo6bo3b2o14b2o3bo6bobo36bo47b2obo29bob2o$b2o8b3o16b3o8b2o
34b3obo45b2o35b2o$74bobobo3bo44bo3b3o25b3o3bo$4bo6bo20bo6bo34bo7bo45bo
b4o25b4obo$3b3o4b3o8b2o8b3o4b3o33b2o4b2o47b3o29b3o$3b2o4b4o8b2o8b4o4b
2o35b2o52bo31bo$9b3o20b3o$9b2o2b2o14b2o2b2o$4b2o7b3o12b3o7b2o$3b2o9bo
6b2o6bo9b2o$2b2o7bob2o6b2o6b2obo7b2o$2b4o5b3o16b3o5b4o$2bob2o6bo18bo6b
2obo$bo2b2o32b2o2bo$2b3o34b3o$3bo36bo$21b2o$21b2o96b2o51b2o$119b2o51b
2o$12bo2bo12bo2bo$12bo2bo12bo2bo$12bo2bo5b2o5bo2bo92bo43bo$21b2o57bo
43b2o41b2o$79b2o40b5o41b5o$16bo10bo50b2o2bo33b2ob2obob2o41b2obob2ob2o$
15bo12bo48b6o33bo7b2o41b2o7bo$80bobo33bo5b4o41b4o5bo$16b2o8b2o51b2o36b
o6b2o6b2o25b2o6b2o6bo$15b2o10b2o90bobo3bo6b3o23b3o6bo3bobo$14bo2bo8bo
2bo90b5o6b4o23b4o6b5o$16b3o6b3o92b3obo5bob2ob2o19b2ob2obo5bob3o$14b2ob
3o4b3ob2o90bo2bo4b4o2b4o17b4o2b4o4bo2bo$13b3ob3ob2ob3ob3o90bobo4b4obo
2b2o17b2o2bob4o4bobo$13b2o2b2o2b2o2b2o2b2o98b2obob2o21b2obob2o$13b2o2b
o8bo2b2o100b2obo23bob2o$133bobo6b2o5b2o6bobo$132b4o5bo2bo3bo2bo5b4o$
131bo3b2o5bobo3bobo5b2o3bo$15b2o10b2o103bobo9b2ob2o9bobo$15b2o10b2o
103b2o2b3o5bo3bo5b3o2b2o$133b2o2b2o15b2o2b2o$11b2o4bo8bo4b2o99bo2b3o
17b3o2bo$11b2o2bo2b2o4b2o2bo2b2o100bo3b2o15b2o3bo$11b2o18b2o102b3o17b
3o$6b3o9bob4obo9b3o$5bo2bo8b4o2b4o8bo2bo$8b3o4bob2o6b2obo4b3o$10bo4b4o
6b4o4bo47bobo$5bo4bo5bo10bo5bo4bo38b3obo2bo$6bo2bo24bo2bo43bobo3$78bo$
12b2o16b2o45b3o$10b5o14b5o$9bo24bo43bo$8bo3bo2bo12bo2bo3bo42bo2bo$9bo
2b2o16b2o2bo44bobo$8bo2bob2o14b2obo2bo45bo44b2o5b3o21b3o5b2o$9bobobo
16bobobo91b2o4b2obo21bob2o4b2o$12bo18bo93bo2bobo6bo17bo6bobo2bo$124b2o
b7o25b7ob2o$123bo5bobo6bo15bo6bobo5bo$123bob3o3bo3bo2bo15bo2bo3bo3b3ob
o$3bo36bo90bo3b3o17b3o3bo$3b3o32b3o88b3o4bo19bo4b3o$4b2o7b3o12b3o7b2o
38bo54bo2bo19bo2bo$4b2o6bo2bo12bo2bo6b2o39bo52b3o23b3o$15bo12bo48b3o
51bob2o2bo17bo2b2obo$3bo2bo4bo20bo4bo2bo37b2o9b3o34bo5b2ob2o19b2ob2o5b
o$3b2o2bo2bo2bo16bo2bo2bo2b2o38bobo6bo2bo34b2o4b2o2bo19bo2b2o4b2o$4b2o
bo3bobo16bobo3bob2o39b3o9bo34b2o5b2obo19bob2o5b2o$5b2o30b2o41b2o5bo46b
4o17b4o$13b3o12b3o55bo2bo33b2o43b2o$4b3o5bo2bo12bo2bo5b3o37b2o8bobo34b
2o41b2o$7bo7bo12bo7bo41b2o45b2o39b2o$11bo3bo12bo3bo46b2o8b3o31b4o7b3o
19b3o7b4o$3bo4bo2b2o3bo10bo3b2o2bo4bo36b4o7bo2bo31b2obo7bo2bo17bo2bo7b
ob2o$bob3obo3bo20bo3bob3obo34b2obo10bo31b2o2bo6bo23bo6bo2b2o$bo10bo2bo
12bo2bo10bo34b2o2bo5bo3bo32b3o7bo3bo15bo3bo7b3o$3b2o7b3o14b3o7b2o37b3o
6b2o3bo32bo7bo3b2o15b2o3bo7bo$2bob2o32b2obo37bo7bo50bo15bo$2bo2bo32bo
2bo46bo2bo42bo2bo17bo2bo$3b3o32b3o47b3o44b3o17b3o!

[Nathaniel]

Kk, got a couple bugfixes uploaded today. The "0 soups" bug is gone and the front page is working again.

[Lewis]

There's a new methuselah, in second place on the table for the standard Life rules (32,370 generations) which needs to be verified. I don't know if a low-population methuselah can be made from it.

[calcyman]

There's a new methuselah, in second place on the table for the standard Life rules (32,370 generations)

I make it 31 319 generations. Interestingly, it collides two gliders to form a pond, then converts it to a block, then finally into a pi-heptomino. Similar to the Chapman-Greene slow salvo technology.

[Lewis]

Was this puffer already known, and if not, is it the first known infinite growth pattern in this rule?

As far as I know, this is the first infinite growth in b35/s23.

[Sokwe]

A pattern was found recently by Richard Wobus that lasts for over 32000 generations (currently second in the list of long-lived patterns). Amazingly, the pattern essentially becomes three blinkers and a B-heptomino, giving rise to this 15-cell 32829-generation methuselah:

Code: Select all

#C 15-cell 32829-generation methuselah
#C found by Richard Wobus with Nathaniel Johnston's soup search (Oct. 7, 2010)
x = 23, y = 31, rule = B3/S23
3o20$3bo$3bo$3bo5$20b3o$9b3o10bo$22bo$21bo!

In terms of its initial population, this is currently the smallest known methuselah with a lifespan over 30000 (being only two cells larger than Lidka).

Edit: Here's a 16-cell 29937-generation methuselah based on a pattern by 'Kang Seonghoon aka senokay' (currently sixth in the list of long-lived patterns):

Code: Select all

x = 23, y = 21, rule = B3/S23
o$2o10b3o$o10bo2bo7$8bo$8b2o$8bo7$22bo$22bo$22bo!

[alan2here]

Given that.

x = 8, y = 8, rule = B36/S125
5bo$4b3o$6b2o$6bo$5bo2$4bo$3bo!

Shouldn't the bounding box for B36/S125 for valid methuselahs be less that 8*8?

[Nathaniel]

Given that.

x = 8, y = 8, rule = B36/S125
5bo$4b3o$6b2o$6bo$5bo2$4bo$3bo!

Shouldn't the bounding box for B36/S125 for valid methuselahs be less that 8*8?

I don't see why. There are patterns in regular Life with bounding box 5x5 that grow indefinitely (see here). A methuselah, however, is a pattern that stays takes a long time to stabilize into predictable behaviour.

[alan2here]

wow, that 10 cell puffer starts off impressively small.

Thanks. So as soon as oscillators, still lives and the various sorts of ships are the only things in existance then you stop counting.

It could be worth trying to make a sort of machine. A binary counter could have a verry long period before it repeated.

Presumably spacefillers with chaotic elements and chaotic (dirty) puffers are predictable as at least dirty puffers eventually repeat.

[137ben]

That definition has always been vague. A methuselah has to be "small", but it is unclear whether this should be done by bounding box or by initial population. It is also not entirely decided whether infinite growth patterns that take awhile to become predictable should be counted, since they never stabilize.
Now, if we want to define how small something has to be, it makes sense to require that a methuselah be smaller than the smallest pattern which can take an arbitrarily long time to stabilize. In B3/S23, this would mean allowing no more than 7 cells (at 8 cells a glider flying towards a blinker an arbitrary distance away can take an arbitrarily long time). Or, instead, we could require a specific sized bounding box, or we could use a combination of bounding box and initial population.

[Sokwe]

Here's another long-lived pattern with an underestimated lifespan (estimated lifespan: 26380 - actual lifespan: 29079):

Code: Select all

x = 20, y = 20, rule = B3/S23
4bo2b2o4b2o$o3bo4bo3bo4bo$3bobo2b2obob2o$b2obob3obob4obo$2obobobobo3bo
5bo$3bob3o2bo$obo3bobobobo$2bo3bo4b4o2bo$2bo2b2ob3o4bo3bo$obo2b3o2b4o
5bo$o4bo3bobo5bo$bob2o8bobo$5bobo3bo4bo2bo$4bobobobo2b4o$bo4b3o3bo3bo
2bo$2ob3ob2o4bo5bo$obobo5bob2o3b2o$o3bo6bo5b2o$8bo2bo7bo$o2bo2bo4b3o3b
obo!

Also, I would encourage Richard Wobus to name the following pattern (mentioned in an earlier post):

Code: Select all

#C 15-cell 32829-generation methuselah
#C found by Richard Wobus with Nathaniel Johnston's soup search (Oct. 7, 2010)
x = 23, y = 31, rule = B3/S23
3o20$3bo$3bo$3bo5$20b3o$9b3o10bo$22bo$21bo!

[knightlife]

Also, I would encourage Richard Wobus to name the following pattern (mentioned in an earlier post):

Code: Select all

#C 15-cell 32829-generation methuselah
#C found by Richard Wobus with Nathaniel Johnston's soup search (Oct. 7, 2010)
x = 23, y = 31, rule = B3/S23
3o20$3bo$3bo$3bo5$20b3o$9b3o10bo$22bo$21bo!

I was able to push it over the top by adding 3 cells:

Code: Select all

#C 18-cell 37673-generation methuselah
#C based on Richard Wobus' 15-cell marvel
x = 75, y = 54, rule = B3/S23
73b2o$73bo22$3o20$3bo$3bo$3bo5$20b3o$9b3o10bo$22bo$21bo!

The bounding box is 75 x 54 but so far, 37673 may be a record for any methuselah of size 18 cells or less (even allowing a 100 x 100 bounding box).
The 15-cell pattern found by Richard Wobus is truly a standout methuselah.

I suppose the soup search will soon pass the 40000 generation barrier.

[Richard]

Also, I would encourage Richard Wobus to name the following pattern (mentioned in an earlier post):

Code: Select all

#C 15-cell 32829-generation methuselah
#C found by Richard Wobus with Nathaniel Johnston's soup search (Oct. 7, 2010)
x = 23, y = 31, rule = B3/S23
3o20$3bo$3bo$3bo5$20b3o$9b3o10bo$22bo$21bo!

If it's up to me, we can name it "Sedna" after the very distant minor planet, which is named after the Inuit sea goddess.

Since I've never really introduced myself here -- I was one of the loyal readers of Martin Gardner's "Mathematical Games" column 40 years ago when Conway's Life was introduced. I was about to be drafted into the army, with no computer access for the next two years, so I played with the "r-pentomino" and other patterns with pennies and nickels or paper and pencil. Even then I realized that this was more interesting than the average "Games" column, and that we would be reading about Life for many years. It was while learning about Golly and the Caterpillar, both of which go well beyond what most of us anticipated in 1970, that I found out about the soup search.

[Mats]

The bounding box is 75 x 54 but so far, 37673 may be a record for any methuselah of size 18 cells or less (even allowing a 100 x 100 bounding box).

This ark found by Nick Gotts and retreived from Stephen Silver's Life Lexicon is well within those limits. It stabilizes at generation 736692...

Code: Select all

x = 32, y = 29, rule = B3/S23
27bo$28bo$29bo$28bo$27bo$29b3o20$2o$2bo$2bo$3b4o!

[137ben]

@above: Arks fall into a different category, as it never actually stabilizes. A more appropriate term would be that it regularizes at time 736692.

[Mats]

Well... A "normal" methuselah doesn't really stabilize eiter. It's bounding box for generation n is expanding forever. That's not really stable. So it's a matter of definition...