Random Glider Collision Search Results

[simsim314]

Following this topic viewtopic.php?f=9&t=1504, and similar to the soup search results thread, this thread is made to share Random Glider Collision Search results.

I will start with this:

Code: Select all

x = 44, y = 31, rule = B3/S23
3b2o$3b2o4$2o27b2o$2o26bo2bo$19bo9bobo$18bobo9bo$18bobo$19bo6$42b2o$
38b2o2b2o$38b2o3$7bo$7b2o$6bobo5$6b2o$5bobo$7bo!

And the initial soup is:

Code: Select all

x = 749, y = 691, rule = B3/S23
76$638bo$637bo$637b3o14$625bo$624bo$624b3o6$577bo$577bobo$577b2o13$
591bo$589b2o$590b2o19$538bobo$538b2o$539bo12$552bo$550b2o$551b2o18$
542bo$540b2o$541b2o8$518bo$518bobo$518b2o13$476bobo$476b2o$477bo12$
461bo$460bo$460b3o6$218bo$219b2o$218b2o2$458bobo$458b2o$459bo$196bo$
197b2o$196b2o10$481bo$479b2o$480b2o5$221bo$222bo$220b3o5$226bobo$227b
2o$227bo4$431bo$429b2o$430b2o9$223bo$221bobo$222b2o4$423bo$423bobo$
423b2o12$265bo$263bobo163bo$264b2o163bobo$429b2o11$271bobo$272b2o$272b
o3$365bo$282bo81bo$280bobo81b3o$281b2o4$316bo$317b2o$316b2o5$280bo$
281bo77bobo$279b3o77b2o$360bo68$297b2o79b2o$298b2o77b2o$297bo81bo6$
382b2o$382bobo$382bo5$405b2o$280bo123b2o$280b2o124bo$279bobo11$275b3o$
277bo$276bo5$239b3o$241bo162b3o$240bo163bo$405bo14$235b2o$234bobo$236b
o2$433b3o$433bo$434bo6$217bo$217b2o$216bobo$459b2o$458b2o$460bo16$197b
2o$196bobo$198bo16$178b3o$180bo$179bo8$187b2o$188b2o$187bo15$164b3o$
166bo$165bo3$179b3o$181bo$180bo4$179b2o$180b2o$179bo13$156bo$156b2o$
155bobo13$141b2o$140bobo$142bo11$128b2o$127bobo$129bo16$107b3o$109bo$
108bo9$92b2o$91bobo$93bo19$90bo$90b2o$89bobo8$47bo$47b2o$46bobo5$46b2o
$45bobo$47bo!

[simsim314]

Here is very interesting 25 bitter:

Code: Select all

x = 540, y = 521, rule = B3/S23
45bo$46bo$44b3o7$33bobo$34b2o$34bo17$56bo$57bo$55b3o9$93bo$94bo$92b3o
18$118bo$116bobo$117b2o16$137bo$138b2o$137b2o5$146bo$144bobo$145b2o10$
136bo$137bo$135b3o18$170bobo$171b2o$171bo6$158bo$159bo$157b3o8$369bobo
$369b2o$370bo4$171bo$172b2o$171b2o3$393bobo$393b2o$394bo6$382bobo$382b
2o$383bo2$200bo$201b2o$200b2o13$239bo$240b2o103bo$239b2o102b2o$344b2o
5$241bobo$242b2o$242bo4$343bobo$252bo90b2o$250bobo91bo$251b2o4$301bobo
$301b2o$302bo5$220bo106bo$221b2o102b2o$220b2o104b2o69$257b3o69b3o$259b
o69bo$258bo71bo6$244b3o$246bo$245bo67b3o$313bo$314bo12$225b2o$226b2o$
225bo135b2o$360b2o$362bo2$214b3o$216bo$215bo4$355b2o$217b2o135b2o$218b
2o136bo$217bo6$185b3o$187bo$186bo2$349b2o$349bobo$172b2o175bo$171bobo$
173bo4$367b2o$367bobo$367bo6$371b2o$370b2o$372bo3$187b2o$186bobo$188bo
6$379b2o$379bobo$379bo11$156b2o$157b2o$156bo2$420b2o$420bobo$420bo6$
149bo$149b2o$148bobo9$434b3o$434bo$435bo3$428b3o$428bo$429bo3$109bo$
109b2o$108bobo6$469b2o$469bobo$469bo7$106b2o$107b2o$106bo363bo$469b2o$
469bobo9$110bo$110b2o338b2o$109bobo338bobo$450bo10$98bo$98b2o$97bobo5$
480b2o$480bobo$480bo4$43b3o$45bo$44bo2$481b2o$480b2o$482bo3$529b2o$
529bobo$529bo$69b2o$68bobo$70bo2$530bo$529b2o$529bobo3$523bo$522b2o$
522bobo9$7b2o$6bobo$8bo7$538b2o$537b2o$539bo11$3o$2bo$bo11$15b2o$16b2o
$15bo!

And here it's reduced:

Code: Select all

x = 10, y = 22, rule = B3/S23
5bo$5bo$5bo2$b3o3b3o2$5bo$5bo$3o2bo7$8bo$7bobo$7b2o2$b2o$2b2o$bo!

Feels like magic.

EDIT In the same manner:

Code: Select all

x = 12, y = 12, rule = B3/S23
b2o$o2bo$o2bo$b3o2$7b3o2$5bo5bo$5bo5bo$5bo5bo2$7b3o!

EDIT2 This looks usable:

Code: Select all

x = 24, y = 17, rule = B3/S23
22b2o$22b2o4$17b3o$19bo$11b2o4b3o$11b2o3$20b2o$bo9b2o6bobo$o4bo5b2o5b
3o$obo3bo12b2o$b3o2bo$2b3o!

and this:

Code: Select all

x = 31, y = 34, rule = B3/S23
3$8bo$8bo$8bo2$19bo$19bo$19bo3b2o$23b2o10$15b3o$15bobo$15bobo2$14bo$
13bobo$13bo2bo$14b2o!

[simsim314]

This one looks hard to synthesize otherwise:

Code: Select all

x = 40, y = 40, rule = B3/S23
5$27b2o$27b2o$19b2o$19b2o7$5b2o$5b2o3$37b3o$37bo$38bo$7b2o$6bobo$8bo
13$12b2o$11bobo$13bo!

[simsim314]

Here is something I've made for fun:

Code: Select all

x = 68, y = 68, rule = B3/S23
60bo$32b2o25bo$32bobo24b3o$33b2o$29bo$28bobo$28bo2bo$29b2o$62b2o$61bo
2bo$61bobo$62bo$66b2o$65bobo$65b2o27$58b2o$57bo2bo$58bobo$59bo$54b2o$
54bobo$55b2o18$b2o$obo$2bo!

[simsim314]

Here is natural synth of Trans-block and long hook eating tub

Code: Select all

x = 460, y = 449, rule = B3/S23
437bobo$437b2o$438bo5$458bo$457bo$457b3o7$427bo$425b2o$426b2o7$437bo$
436bo$436b3o6$428bo$428bobo$428b2o9$431bo$431bobo$431b2o16$390bobo$
390b2o$391bo3$414bo$414bobo$414b2o15$389bobo$389b2o$390bo11$337bo$337b
obo$obo334b2o$b2o$bo11$338bo$337bo$337b3o$bo$2bo$3o6$21bobo$22b2o$22bo
3$346bo$345bo$345b3o6$343bobo$17bo325b2o$18bo325bo$16b3o13$25bo$26b2o$
25b2o2$301bo$301bobo$301b2o5$311bo$309b2o$51bobo256b2o$52b2o$52bo10$
82bo219bo$80bobo219bobo$81b2o219b2o2$287bo$286bo$43bo242b3o$44bo$42b3o
17$110bo167bo$111bo165bo$109b3o165b3o8$83bo$81bobo$82b2o6$218bo$218bob
o$218b2o9$116bo84bo$114bobo84bobo$115b2o84b2o5$127bo$125bobo$126b2o2$
231bobo$231b2o$232bo7$229bo$111bobo114bo$112b2o114b3o$112bo66$127b2o$
128b2o84b3o$127bo86bo$215bo3$142b3o$144bo$143bo6$145bo$145b2o$144bobo
4$211b2o$90b3o117b2o$92bo119bo$91bo3$245b2o$244b2o$246bo7$222b2o$221b
2o$223bo4$86b2o$87b2o$86bo7$251b3o$251bo$105b3o144bo$107bo$106bo12$73b
2o$72bobo$74bo3$291b3o$291bo$70b3o219bo$72bo$71bo11$27bo$27b2o$26bobo
18$21b3o$23bo$22bo18$34b3o$36bo$35bo!

Some 32 bitter:

Code: Select all

x = 502, y = 421, rule = B3/S23
8bo$6bobo$7b2o9$5bo$6bo$4b3o13$2bo$obo$b2o10$28bo$29b2o$28b2o20$47bobo
$48b2o$48bo14$80bo$78bobo$79b2o16$108bo$109b2o$108b2o7$488bo$487bo$
487b3o8$89bo$90bo$88b3o4$499bo$499bobo$499b2o3$125bobo$126b2o$126bo8$
468bo$466b2o$467b2o8$125bo$126b2o$125b2o6$149bo304bo$147bobo303bo$148b
2o303b3o10$167bo$165bobo$166b2o6$422bo$422bobo$422b2o6$411bobo$411b2o$
412bo4$158bobo$159b2o$159bo4$407bo$407bobo$407b2o2$168bo$169bo$167b3o
9$400bo$400bobo$192bo207b2o$190bobo$191b2o7$226bobo$227b2o$227bo6$367b
o$366bo$366b3o7$246bo$247bo$245b3o4$389bo$389bobo$389b2o6$222bo$220bob
o$221b2o3$378bo$377bo$377b3o11$265bo$263bobo$264b2o3$344bobo$344b2o$
345bo3$311bo$282bobo25bo$283b2o25b3o$283bo64$345bo$284b2o58b2o$285b2o
57bobo$284bo3$252b2o$253b2o$252bo9$330b2o$330bobo$330bo$238b2o$237bobo
$239bo2$341b2o$340b2o$342bo9$237b2o$236bobo$238bo125b2o$363b2o$365bo6$
211b2o$210bobo$212bo7$212b3o198b2o$214bo198bobo$213bo199bo17$409b2o$
204b2o202b2o$205b2o203bo$204bo!

[simsim314]

I wonder if this can be used to optimize the synth of 60P5H2V0.

Code: Select all

x = 21, y = 33, rule = B3/S23
5$10bo$9bobo$9bobo$10bo2$5b2o7b2o$4bo2bo5bo2bo$5b2o7b2o6$10bo$9b3o$8b
2ob2o$9b3o$10bo4$9b2o$9b2o!

as the base, instead of this:

Code: Select all

x = 57, y = 51, rule = B3/S23
23bo$24bo$22b3o21bo$44b2o$45b2o3bobo$15bo34b2o$13bobo8bo26bo$14b2o6bob
o$23b2o$38bo13bo$38bobo10bo$38b2o11b3o3$8bo$9b2o$8b2o$37bo$36bobo10bob
o$27b2o6bo2bo10b2o$27bobo6b2o12bo$24b2obobo$24b2obob2o8b3o$28bo3bo2bo$
29b7o2$29b7o$28bo3bo2bo$b2o21b2obob2o$obo21b2obobo$2bo24bobo6b2o$9b2o
16b2o6bo2bo15b2o$10b2o24bobo15bobo$9bo27bo16bo6$38b2o$38bobo$38bo$23b
2o$14b2o6bobo$13bobo8bo26bo$15bo34b2o$45b2o3bobo$44b2o$22b3o21bo$24bo$
23bo!

[dvgrn]

I wonder if this can be used to optimize the synth of 60P5H2V0.

Code: Select all

x = 21, y = 33, rule = B3/S23
5$10bo$9bobo$9bobo$10bo2$5b2o7b2o$4bo2bo5bo2bo$5b2o7b2o6$10bo$9b3o$8b
2ob2o$9b3o$10bo4$9b2o$9b2o!

It seems possible, but maybe not too likely (not that I know what I'm talking about). There's a reason why the stable wings are constructed at the front before the tail part is constructed. It seems as if it would be difficult, and therefore probably expensive, to turn on the red cells in the diagram below:

Code: Select all

x = 12, y = 13, rule = LifeHistory
7.2B$6.B.B$6.B.B.2B$5.DCAB.2B$C2.C3.C$7C2$7C$C2.C3.C$5.DCAB.2B$6.B.B.
2B$6.B.B$7.2B!

Those cells can't turn on unless two other temporary ON cells appear around them, and every such combination seems to cause destructive deaths or births nearby in the next few ticks, threatening the stability of the base object. By the time that damage has been patched up -- if that can be done at all -- there might not be much left of the base still life, and the old synthesis might end up being cheaper.

[As usual when I offer this kind of crackpot opinion, I'm really just inviting someone to prove me wrong...]

[Extrementhusiast]

I wonder if this can be used to optimize the synth of 60P5H2V0.

Code: Select all

x = 21, y = 33, rule = B3/S23
5$10bo$9bobo$9bobo$10bo2$5b2o7b2o$4bo2bo5bo2bo$5b2o7b2o6$10bo$9b3o$8b
2ob2o$9b3o$10bo4$9b2o$9b2o!

It seems possible, but maybe not too likely (not that I know what I'm talking about). There's a reason why the stable wings are constructed at the front before the tail part is constructed. It seems as if it would be difficult, and therefore probably expensive, to turn on the red cells in the diagram below:

Code: Select all

x = 12, y = 13, rule = LifeHistory
7.2B$6.B.B$6.B.B.2B$5.DCAB.2B$C2.C3.C$7C2$7C$C2.C3.C$5.DCAB.2B$6.B.B.
2B$6.B.B$7.2B!

Those cells can't turn on unless two other temporary ON cells appear around them, and every such combination seems to cause destructive deaths or births nearby in the next few ticks, threatening the stability of the base object. By the time that damage has been patched up -- if that can be done at all -- there might not be much left of the base still life, and the old synthesis might end up being cheaper.

[As usual when I offer this kind of crackpot opinion, I'm really just inviting someone to prove me wrong...]

No, I think you're absolutely right, at least for the moment. I recently came up with part of a method to do something similar, but it's ugly, expensive, and the old method is cheaper, anyway:

Code: Select all

x = 190, y = 27, rule = B3/S23
91bo$89bobo$90b2o$2bo90bobo4bo$3bo89b2o4bo$b3o90bo4b3o$111b3o3bo$11b2o
3bo16bo6b2o3bo20b2o3bo20b2o3bo15bob2o10b2o29b2o25b2o$11bobobobo16bo5bo
bobobo19bobobobo19bobobobo13bo3b2o8bobo28bobo24bobo$7b2o4bobobo14b3o2b
o4bobobo21bobobo21bobobo27bobob2o25bobob2o21bobob2o$6bobo4bobob2o17bob
o2b2obob2o15b2o2b2obob2o15b2o2b2obob2o21b2o2b2obob2o24b2obob2o20b2obob
2o$6bobobo3bo17b3o2bo5bo19bo5bo19bo5bo18b2o5bo5bo23bo2bo3bo19bo2bo3bo$
b4o2b2ob4o20bo3b5o21b5o21b5o18bobo6b5o24b7o20b7o$o3bo28bo25bo55bo30bo$
4bo2b2ob4o24b5o17bo3b5o21b5o27b5o17bobo6b5o20b7o$o2bo2bobobo3bo22bo5bo
14b3o2bo5bo19bo5bo25bo5bo17b2o5bo5bo19bo2bo3bo$6bobo4bobob2o17bobo2b2o
bob2o14bobo2b2obob2o15b2o2b2obob2o21b2o2b2obob2o20b2o2b2obob2o20b2obob
2o$7b2o4bobobo19bo4bobobo11b3o2bo4bobobo21bobobo27bobob2o25bobob2o21bo
bob2o$11bobobobo22bobobobo13bo5bobobobo19bobobobo27bobo14bo3b2o8bobo
24bobo$11b2o3bo23b2o3bo13bo6b2o3bo20b2o3bo29b2o15bob2o10b2o25b2o$142b
3o3bo$b3o90bo4b3o$3bo89b2o4bo$2bo90bobo4bo$90b2o$89bobo$91bo!

I don't think there's a cheap synthesis for a long bookend with sesquihat with integral on the proper side.

[simsim314]

Maybe this one can be better base for snark synth:

Code: Select all

x = 12, y = 16, rule = B3/S23
10b2o$4bo5b2o$3bobo$4bo4$4b3o$4bo2b2o$5b3o$6bo4$2ob2o$2ob2o!

EDIT

Or this one:

Code: Select all

x = 22, y = 39, rule = B3/S23
b2o$o2bo$obo$bo2$11b3o$11bobo$11b3o$b2o3b2o$obo3bobo$o7bo$o3bo3bo$bo5b
o$2bo3bo$3b3o22$19b3o$19bo$20bo!

[simsim314]

This might be an improved path for snark-synth:

Code: Select all

x = 67, y = 22, rule = B3/S23
15bo$15bo$15bo2$11b3o3$7bo$5b2ob2o$5b2ob2o25bo$35b3o24bo$38bo22bobo$3o
9b2o21b4o7bobobo10bobo2bo$o11b2o20bo25b2ob4o$3o31bobo25bo$35b2o25bobo$
63b2o4$2b2o$2b2o!

[simsim314]

Tim's Coe p8 synth (slow salvo friendly):

Code: Select all

x = 39, y = 34, rule = B3/S23
11$23bo$23bo$23bo5$20b3o$19bo3bo$18bo5bo$18bo5bo$18bo5bo$19bo3bo4bobo$
20b3o4bo2bo$27bo2bo$28b2o2$12b2o3b2o$12b2o3b2o!

[codeholic]

How do you see it's slow salvo friendly?

[simsim314]

It requires only two synchronized gliders for ignition. This makes the dried salvo pretty small.

[codeholic]

Ah, if you look at it that way...

[simsim314]

This one could make the boojum reflector usable in circuitry:

Code: Select all

x = 15, y = 9, rule = B3/S23
4bo$3bobo$3b2o2$2o$2o10bobo$4bo6bo2bo$4bo6bo2bo$4bo7b2o!

[dvgrn]

This one could make the boojum reflector usable in circuitry...

Not quite, unfortunately, unless you have an easy slow-salvo-compatible way to turn one of the two blocks into an aircraft carrier or something equivalent -- three cells, disconnected on one side. An eater will work, but a boat or ship won't.

We do have Kazyan's 10-glider construction of an eater2 based on an apgsearch soup, which is really eight gliders plus a block, or maybe seven gliders plus a block and a pre-traffic light.

So it's certainly not impossible to build a seed for a boojum reflector -- but it's expensive enough that it's generally possible to design cheaper pure-Spartan Herschel circuitry to perform the same task. Some of Paul Callahan's old stable reflectors were more or less Spartan. The following one does the same thing as a boojum reflector, and offers optional output gliders and/or a Herschel as a bonus:

Code: Select all

#C Paul Callahan's Spartan stable reflector, October 1996
x = 124, y = 97, rule = B3/S23
87b2o$87b2o3$94b2o$94b2o21$84b2o$84b2o7b2o$93bo$91bobo$91b2o$75b2o$76b
o$76bobo$77b2o2$93bo$91b3o$90bo11bo$75bo14b2o8b3o$74bobo22bo$75bo23b2o
14b2o$115b2o3$121b2o$121b2o$117b2o$117b2o3$86b2o$87bo34b2o$84b3o35b2o$
84bo2$bo$2bo$3o$68b2o$67bobo16b2o$67bo19bo$66b2o19bobo$88b2o2$17b2o$
17b2o52bo$55bo15b3o$53b3o18bo44b2o$52bo20b2o11b2o31bo$52b2o32b2o32b3o$
43b2o77bo$43bobo$44bo3$55b2o37b2o21b2o$24b2o29b2o36bobo21b2o$24b2o67bo
17b2o$92b2o17b2o$6b2o$5bobo$5bo107b2o$4b2o100b2o5b2o$65b2o39b2o$66bo$
63b3o20b2o$15b2o46bo22bo$14bobo70b3o$14bo74bo$13b2o$26b2o$26b2o3$28b2o
$28b2o!

This would probably be about the same slow-construct cost as a seven-glider synchronized eater2 seed.

EDIT: It may not be widely known that there is a Spartan version of the boojum reflector. Sort of, anyway. It just needs an extra reset glider from an awkward direction after every use. Should really be posted in Useless Discoveries, no doubt, but while we're on the subject...

Code: Select all

#C Boojum reflector with Spartan reset instead of eater2
x = 95, y = 57, rule = B3/S23
6bo6b2o30bo$7bo5b2o29bobo$5b3o37bo3$39b2o$39b2o23b2o$31b2o31bo$32bo29b
obo$32bobo27b2o$33b2o4b2o25b2o$39b2o25b2o$20b2o$20b2o$38b2o$2b2o34b2o$
bobo$bo$2o51b2o$53b2o$34b2o$34b2o4b2o$11b2o27bobo$10bobo29bo$10bo31b2o
$9b2o23b2o$34b2o3$29bo$28bobo29b2o$29bo30b2o23$92b3o$92bo$93bo!

EDIT2: Of course the same is true of the rectifier, which is a much simpler pattern with only four still lifes. It just occurred to me that a bait still life could also possibly be rebuilt at the key location, to do an asynchronous reset:

Code: Select all

#C rectifier with an output signal to rebuild a 2sL bait
x = 42, y = 31, rule = B3/S23
bobo7bo$2b2o6bobo$2bo7bobo$11bo12$20b2o$20b2o2$2b2o$bobo$bo$2o$31b2o$
30bo2bo$30bobo$11b2o18bo$10bobo21b2o$10bo23bobo$9b2o25bo3bo$36b2obobo$
40bo!

[Kazyan]

We do have Kazyan's 10-glider construction of an eater2 based on an apgsearch soup, which is really eight gliders plus a block, or maybe seven gliders plus a block and a pre-traffic light.

Not that it helps with slow-salvo construction, but I feel the need to point out my newer 6-glider method for the eater2. It's much prettier.

[dvgrn]

Not that it helps with slow-salvo construction, but I feel the need to point out my newer 6-glider method for the eater2. It's much prettier.

Wow -- I missed that posting somehow. Very nice!

I think it does help with slow-salvo construction, actually. Seeds for glider recipes seem to get exponentially more expensive the more gliders you have to synchronize, so this recipe is many times cheaper, not just 33% cheaper. Now Paul Callahan's big old October 1996 device is pretty definitely more expensive to slow-construct than a boojum reflector... though someone will have to actually wire up an eater2 seed before we'll know for sure.

Looks as if there are a couple of pairs of crashing gliders that produce active reactions that might possibly be made by crashing one glider into a small ash object instead (?). Swapping those out would increase the standard glider-recipe cost, of course, but might decrease the slow-salvo construction cost; it's better to need only four or five synchronized gliders at the final stage instead of six.

[simsim314]

Not quite, unfortunately, unless you have an easy slow-salvo-compatible way to turn one of the two blocks into an aircraft carrier or something equivalent -- three cells, disconnected on one side.

Definitely - I posted stage one, and you're talking about stage two. If we manage to find such converter with reasonable amount of seeds, we got a boojum for less than 20 SLs. Not cheap, but it works faster, and would compete with usual spartan circuitry.

[Kazyan]

Looks as if there are a couple of pairs of crashing gliders that produce active reactions that might possibly be made by crashing one glider into a small ash object instead (?). Swapping those out would increase the standard glider-recipe cost, of course, but might decrease the slow-salvo construction cost; it's better to need only four or five synchronized gliders at the final stage instead of six.

Good to know it's helpful. One of the collisions does something pretty elementary, so it can be replaced:

Code: Select all

x = 33, y = 17, rule = B3/S23
11bo$9b2o$10b2o$2bo$obo$b2o3$5bobo$5b2o21bo$6bo20bobo$27bobo$17bo8b2ob
3o$15b2o15bo$5b2o5b2o2b2o8b2ob3o$6b2o4b2o12b2obo$5bo!

...but the others look immutable. The B-heptomino predecessor used is pretty specific to that two-glider collision, and the other gliders have timing issues.