Smaller sawtooth

For discussion of specific patterns or specific families of patterns, both newly-discovered and well-known.
User avatar
Scorbie
Posts: 1692
Joined: December 7th, 2013, 1:05 am

Re: Smaller sawtooth

Post by Scorbie » April 28th, 2015, 4:22 pm

Not Too sure, but I think the sawtooth could be even smaller by using the new p120 gun.

User avatar
biggiemac
Posts: 515
Joined: September 17th, 2014, 12:21 am
Location: California, USA

Re: Smaller sawtooth

Post by biggiemac » April 28th, 2015, 7:06 pm

Here's an appropriate 6hd gun at p120 with 8 fewer cells (119) at minimum population than the twin bees version. If someone can line up the blocker/ship the same as before it could break 200.

Code: Select all

x = 73, y = 71, rule = B3/S23
64b2o$64b2o3$61b2o$61b2o8b2o$5bob2o62bo$5b2obo55b2o6bo$64b2o5b2o3$54bo
$53b3o$2o5b2o10bo36bo$2o5b2o11b2o32b2o$15bobobo35bo$4b2o9bobo$4b2o4$
27b2o$27b2o2$24b2o5b2o20b2o$24b2o5b2o20b2o2$56b2o$56b2o3$23bo29b2o$21b
obo29b2o$22b2o4$30bo$20bo9bobo$20b3o7b2o$23bo$22b2o4$25bo$23b2ob2o2$
22bo5bo2$22b2obob2o9$24b2o$24b2o$53bo$51bobo$52b2o5$53bo$54b2o$53b2o!
Edit: upon further investigation, the blocker won't work if the block is pulled at p8n. The second to last block pull needs to happen with the blocker out of phase or it'll make a mess. So we need a different deletion mechanism. Anything promising at p120?

There's also a 99-cell 10hd 3fd-block-pull mechanism using the double barrel p120 gun, but it isn't compatible with the ship.

Code: Select all

x = 54, y = 59, rule = B3/S23
$23b2o$23bo$11bobo7bobo$11bo2bo6b2o$2b2o10b2o$2b2o8bo3b2o$7b2o5b2o$6bo
4bo2bo$11bobo6$2bo$2b3o$5bo$4b2o$10bo4b2o5b2o10bo$11bo3b2o5b2o11b2o$9b
3o18bobobo$19b2o9bobo$19b2o4$42b2o$4b2o3b2o31b2o$6b3o$5bo3bo29b2o5b2o$
6bobo30b2o5b2o$7bo4$6b2o$6b2o30bo$36bobo$37b2o10$40bo$41bo$39b3o6$52b
2o$52b2o!
Re-edit: Ah, a pd, of course. Nice job chris, and of course excellent find simsim!
Last edited by biggiemac on April 28th, 2015, 7:29 pm, edited 2 times in total.
Physics: sophistication from simplicity.

chris_c
Posts: 966
Joined: June 28th, 2014, 7:15 am

Re: Smaller sawtooth

Post by chris_c » April 28th, 2015, 7:16 pm

biggiemac wrote:Here's an appropriate 6hd gun at p120 with 8 fewer cells at minimum population than the twin bees version. If someone can line up the blocker/ship the same as before it could break 200.
So far I've got a Sawtooth183 that reaches minimum population at generation 108 * (121^n - 1). Kudos to simsim for finding the quite shocking reduction of the p120 gun.

Code: Select all

x = 86, y = 98, rule = B3/S23
5bob2o$5b2obo3$77b2obo$77bob2o$10bo$2o3bo4b2o$2o2b2o5b2o$5b2o4bo$6bo3b
o$77b2o5b2o$77b2o5b2o2$80b2o$27b2o51b2o$27b2o2$24b2o5b2o$24b2o5b2o24b
2o$17bo39b2o12bo$18b2o51bobo$17b2o34b2o5b2o9b3o$53b2o5b2o11bo21$48bo$
46b2o$47b2o$54bo$52b3o$51bo$51b2o$46bo$45bo$45b3o7$46b2o3b2o$48b3o$47b
o3bo$48bobo$49bo4$49b2o$49b2o3$22b2o12bo4bo$34b2ob4ob2o$25bo10bo4bo$
14b4o$12b2o4b2o$12b2o5bo$14b2obobo$19bo$15bo3bo$15bo4bo$17b3o3bo$17b2o
4bo$23b2o$25bo$25b3o3$28bo$27bob5o$26b2o5bo$26b2o3bo2bo$34bo$28b2obo2b
o$31bo2bo$32b2o$32b2o!
EDIT: Now Sawtooth181 with minimum population at 58 * (121^n - 1).

Code: Select all

x = 78, y = 87, rule = B3/S23
5bob2o$5b2obo3$69b2obo$69bob2o$6bo$2o3bo3b3o$2o2bo6bo$5bo5bo$6b2ob2o$
69b2o5b2o$69b2o5b2o2$72b2o$27b2o43b2o$27b2o2$24b2o5b2o$24b2o5b2o16b2o$
18bo30b2o9bobo$16bobo41bobobo$17b2o26b2o5b2o11b2o$45b2o5b2o10bo20$50bo
$39bo8b3o$39bobo5bo$39b2o6b2o3$45bo$44bobo$39bo3bo3bo$37b2o5b3o$38b2o
2b2o3b2o11$17b4o24b2o$15b2o4b2o22b2o$15b2o5bo$17b2obobo$22bo9bo4bo$18b
o3bo7b2ob4ob2o$18bo4bo8bo4bo$20b3o3bo$20b2o4bo$26b2o$28bo$28b3o3$31bo$
30bob5o$29b2o5bo$29b2o3bo2bo$37bo$31b2obo2bo$34bo2bo$35b2o$35b2o!

User avatar
Freywa
Posts: 877
Joined: June 23rd, 2011, 3:20 am
Location: Singapore
Contact:

Re: Smaller sawtooth

Post by Freywa » April 28th, 2015, 8:23 pm

chris_c wrote:So far I've got a Sawtooth183 that reaches minimum population at generation 108 * (121^n - 1). Kudos to simsim for finding the quite shocking reduction of the p120 gun.

EDIT: Now Sawtooth181 with minimum population at 58 * (121^n - 1).
There's actually quite a relation between the period of the gun and the expansion factor. Sawtooth 213/201 had p46 guns and an expansion factor of 46 + 1; sawtooth 183/181 uses p120 guns and has an expansion factor of 120 + 1. Both of them use the (-1, -1) two-glider block retraction.
Princess of Science, Parcly Taxel

Code: Select all

x = 31, y = 5, rule = B2-a/S12
3bo23bo$2obo4bo13bo4bob2o$3bo4bo13bo4bo$2bo4bobo11bobo4bo$2bo25bo!

towerator
Posts: 328
Joined: September 2nd, 2013, 3:03 pm

Re: Smaller sawtooth

Post by towerator » May 2nd, 2015, 1:36 pm

biggiemac wrote: So you've made a 163-cell predecessor to the sawtooth, but still only with a minimum repeating population of 201. Although something goes wrong with your pattern after two sawtooth iterations, it explodes..

Edit: Putting them side by side I see that your pattern deleted the gliders, making the phase different, and ultimately leading to instability.
Strangely, I had made adjustings on the unix to get back to stability... whatever. Please ignore this post.
I thought the record was "Smallest pattern to exhibit sawtooth mechanism".
This is game of life, this is game of life!
Loafin' ships eaten with a knife!

User avatar
simsim314
Posts: 1823
Joined: February 10th, 2014, 1:27 pm

Re: Smaller sawtooth

Post by simsim314 » July 9th, 2015, 6:35 pm

Better later than never...congrats to Chris for finding Sawtooth 181! We should update Wiki.

I was checking Wiki being sure I'll find Chris's sawtooth and couldn't find it...

User avatar
calcyman
Moderator
Posts: 2932
Joined: June 1st, 2009, 4:32 pm

Re: Smaller sawtooth

Post by calcyman » October 25th, 2015, 1:12 pm

Shameless corollary snipe to re-establish myself at the head of the Sawtooth Hall of Fame:

Code: Select all

#C Sawtooth 179
#C Adam P. Goucher, 2015-10-25
x = 72, y = 83, rule = B3/S23
63b2obo$63bob2o6$63b2o5b2o$63b2o5b2o$7b2o$7b2o57b2o$66b2o2$10b2o$2o8b
2o$bo41b2o$o6b2o34b2o9bobo$2o5b2o45bobobo$39b2o5b2o11b2o$39b2o5b2o10bo
$18bo$17b3o$16bo$17b2o$17bo9$18b2o$18b2o2$15b2o$15b2o2$44bo$18b2o13bo
8b3o$18b2o13bobo5bo$33b2o6b2o3$39bo$38bobo$33bo3bo3bo$31b2o5b3o$32b2o
2b2o3b2o11$11b4o24b2o$9b2o4b2o22b2o$9b2o5bo$11b2obobo$16bo9bo4bo$12bo
3bo7b2ob4ob2o$12bo4bo8bo4bo$14b3o3bo$14b2o4bo$20b2o$22bo$22b3o3$25bo$
24bob5o$23b2o5bo$23b2o3bo2bo$31bo$25b2obo2bo$28bo2bo$29b2o$29b2o!
What do you do with ill crystallographers? Take them to the mono-clinic!

User avatar
simsim314
Posts: 1823
Joined: February 10th, 2014, 1:27 pm

Re: Smaller sawtooth

Post by simsim314 » October 26th, 2015, 2:59 am

calcyman wrote:Shameless corollary snipe to re-establish myself at the head of the Sawtooth Hall of Fame
Congrats! We should update wiki once again.

thunk
Posts: 170
Joined: October 3rd, 2015, 8:50 pm
Location: Central USA

Re: Smaller sawtooth

Post by thunk » October 27th, 2015, 11:43 pm

calcyman wrote:Shameless corollary snipe to re-establish myself at the head of the Sawtooth Hall of Fame:

Code: Select all

rle
Well, corollary sniping is fun!

Code: Select all

#C Sawtooth 177
#C thunk, 2015-10-27
#C Pop=177 at T=15, 6975...
x = 68, y = 76, rule = B3/S23
59b2o$59b2o3$56b2o$56b2o8b2o$7b2o57bo$7b2o50b2o6bo$59b2o5b2o2$10b2o54b
2o$2o8b2o55bo$bo61bo$o6b2o53bo3bo$2o5b2o55b2o2$2o$o$4bo43b3o$bo3bo41bo
2bo$2b2o43bo3bo$46b2obobo$46b2ob2o$47b3o$17b3o$17bo2bo$16bo3bo29b2o$
16bobob2o27bo2bo$17b2ob2o25b5o$18b3o25b2ob3o$46b3o$47bobo$16b2o22bo7b
2o$15bo2bo10bo8b3o$16b5o8bobo5bo$16b3ob2o7b2o6b2o$19b3o$18bobo$18b2o
15bo$34bobo$29bo3bo3bo$27b2o5b3o$28b2o2b2o3b2o11$7b4o24b2o$5b2o4b2o22b
2o$5b2o5bo$7b2obobo$12bo9bo4bo$8bo3bo7b2ob4ob2o$8bo4bo8bo4bo$10b3o3bo$
10b2o4bo$16b2o$18bo$18b3o3$21bo$20bob5o$19b2o5bo$19b2o3bo2bo$27bo$21b
2obo2bo$24bo2bo$25b2o$25b2o!
The Simkin guns were off-phase by 2 ticks. Rephasing one allowed both Herschels to be in their smallest canonical 7-cell form, and bringing them 4 cells closer synchronized both with the smallest pentadecathlon form.

EDIT: Fixed second minimum population time.
"What's purple and commutes?
The Evanston Express."

User avatar
Scorbie
Posts: 1692
Joined: December 7th, 2013, 1:05 am

Re: Smaller sawtooth

Post by Scorbie » October 28th, 2015, 2:05 am

If I got it right, all the components are in its minimal phase, so there won't be any more improvements with trivial (i.e. without change of mechanism) collorary sniping, right??

chris_c
Posts: 966
Joined: June 28th, 2014, 7:15 am

Re: Smaller sawtooth

Post by chris_c » October 28th, 2015, 7:07 am

thunk wrote: Well, corollary sniping is fun!

Code: Select all

#C Sawtooth 177
#C thunk, 2015-10-27
Very good. It seemed impossible to me to get the phases matched up like that. Well done! Note that I would suggest the following as the canonical version of your pattern. Population is 177 at 48 * (121^n-1).

Code: Select all

x = 68, y = 77, rule = B3/S23
59b2o$59b2o3$56b2o$56b2o8b2o$7b2o57bo$7b2o50b2o6bo$59b2o5b2o2$10b2o$2o
8b2o$bo46b3o$o6b2o40bo$2o5b2o40b3o4$17b3o$18bo$16b3o4$48b2o$48b2o2$51b
2o$51b2o2$18b2o$18b2o28b2o$40bo7b2o$15b2o21b3o$15b2o20bo$37b2o2$18b2o
5bobo$18b2o5b2o$26bo5$24bo$24bobo5b2o3b2o$24b2o8b3o$33bo3bo$34bobo$35b
o4$35b2o$6b4o25b2o$4b2o4b2o$4b2o5bo$6b2obobo10bo4bo$11bo8b2ob4ob2o$7bo
3bo10bo4bo$7bo4bo$9b3o3bo$9b2o4bo$15b2o$17bo$17b3o3$20bo$19bob5o$18b2o
5bo$18b2o3bo2bo$26bo$20b2obo2bo$23bo2bo$24b2o$24b2o!
It is possible to move the ship closer by one more cell but then the minimum population is only achieved every second cycle because of the boat bits.

User avatar
calcyman
Moderator
Posts: 2932
Joined: June 1st, 2009, 4:32 pm

Re: Smaller sawtooth

Post by calcyman » October 28th, 2015, 10:55 am

thunk wrote:Well, corollary sniping is fun!
My record didn't last very long. Many congratulations to you!
chris_c wrote:Well done! Note that I would suggest the following as the canonical version of your pattern. Population is 177 at 48 * (121^n-1).
I concur; the canonical version should have the minimum expansion factor (121 in this case) and attain its minimum population in generation 0. Subject to those constraints, it's fashionable to minimise the bounding box.

I imagine this record will be extremely difficult to improve further. We either need a smaller shotgun, or a smaller spaceship, or a completely different tractor-beam mechanism.
What do you do with ill crystallographers? Take them to the mono-clinic!

User avatar
dvgrn
Moderator
Posts: 10612
Joined: May 17th, 2009, 11:00 pm
Location: Madison, WI
Contact:

Re: Smaller sawtooth

Post by dvgrn » October 28th, 2015, 11:49 am

calcyman wrote:I imagine this record will be extremely difficult to improve further. We either need a smaller shotgun, or a smaller spaceship, or a completely different tractor-beam mechanism.
Just to make the list complete, a smaller 90-degree reflector is also a theoretical possibility. Any decade now, Catagolue might turn up a new [p10|p12|p15|p18|p20|p24] oscillator with a glider-turning spark, that happens to have less than 23 ON cells in the key phase.

Or CatForce might turn up a five-block constellation that recovers after turning a glider, or maybe it will be a mixed constellation including a blinker or two.

All of these are impressively low-probability, I must admit.

User avatar
Freywa
Posts: 877
Joined: June 23rd, 2011, 3:20 am
Location: Singapore
Contact:

Re: Smaller sawtooth

Post by Freywa » October 28th, 2015, 12:20 pm

thunk wrote:...
Well, thunk, I implore to know what your real name and (optionally) location is for the sake of posterity. Conway's Life is still very much academic in nature.
Princess of Science, Parcly Taxel

Code: Select all

x = 31, y = 5, rule = B2-a/S12
3bo23bo$2obo4bo13bo4bob2o$3bo4bo13bo4bo$2bo4bobo11bobo4bo$2bo25bo!

thunk
Posts: 170
Joined: October 3rd, 2015, 8:50 pm
Location: Central USA

Re: Smaller sawtooth

Post by thunk » October 28th, 2015, 2:11 pm

chris_c wrote:Very good. It seemed impossible to me to get the phases matched up like that. Well done! Note that I would suggest the following as the canonical version of your pattern. Population is 177 at 48 * (121^n-1).
Thanks. :) I would agree that that should be the canonical version, I just wasn't sure what the usual conventions for posting sawtooths were.
Scorbie wrote:If I got it right, all the components are in its minimal phase, so there won't be any more improvements with trivial (i.e. without change of mechanism) collorary sniping, right??
I think so too. The only B60 phases with smaller population are 39 and 40 ticks after the initial Herschel position, and none of those work in the sawtooth--at 40 ticks, there end up being two sets of gliders in the pattern when the block is finally deleted (10 cell penalty), and at 100 ticks, the FNGs have not yet been deleted by the boat-bits.
Freywa wrote: Well, thunk, I implore to know what your real name and (optionally) location is for the sake of posterity. Conway's Life is still very much academic in nature.
Just so you know, demanding to out people in public is rather rude. Many people have reasons for not wanting to divulge their birth name, and I'm not the first (Guam comes to mind) to prefer pseudonymity.

Anyway, a nym didn't prevent this guy from being published.
"What's purple and commutes?
The Evanston Express."

User avatar
calcyman
Moderator
Posts: 2932
Joined: June 1st, 2009, 4:32 pm

Re: Smaller sawtooth

Post by calcyman » October 30th, 2015, 6:40 am

thunk wrote:
Scorbie wrote:If I got it right, all the components are in its minimal phase, so there won't be any more improvements with trivial (i.e. without change of mechanism) collorary sniping, right??
I think so too. The only B60 phases with smaller population are 39 and 40 ticks after the initial Herschel position, and none of those work in the sawtooth--at 40 ticks, there end up being two sets of gliders in the pattern when the block is finally deleted (10 cell penalty), and at 100 ticks, the FNGs have not yet been deleted by the boat-bits.
Yes, I'm pretty certain it is optimal based on present technology. As such, I've created a wiki page for your discovery: http://conwaylife.com/wiki/Sawtooth_177
thunk wrote:Just so you know, demanding to out people in public is rather rude. Many people have reasons for not wanting to divulge their birth name, and I'm not the first (Guam comes to mind) to prefer pseudonymity.
I tend to adhere to Gaetane's advice on this matter, but it's entirely a matter of personal preference: http://security.stackexchange.com/quest ... 6631#46631
thunk wrote:Anyway, a nym didn't prevent this guy from being published.
Or indeed Satoshi Nakamoto.
What do you do with ill crystallographers? Take them to the mono-clinic!

thunk
Posts: 170
Joined: October 3rd, 2015, 8:50 pm
Location: Central USA

Re: Smaller sawtooth

Post by thunk » October 31st, 2015, 1:22 am

Using the other side of the spaceship allows for some bounding box reduction.

Code: Select all

x = 158, y = 74, rule = B3/S23
54b2obo$54bob2o2$12b2o$12b2o$46bo$46b2o$15b2o28bobo6b2o5b2o$5b2o8b2o
37b2o5b2o$6bo$5bo6b2o22bo20b2o$5b2o5b2o23b2o18b2o$36bobo2$149b2o$22b3o
10b2o112b2o$23bo11b3o$21b3o11b2ob2o$30b2o3b2ob2o106b2o$30b2o4bob2o106b
2o8b2o$37bo59b2o57bo$97b2o50b2o6bo$149b2o5b2o2$100b2o$90b2o8b2o$91bo
46b3o$23b2o65bo6b2o40bo$23b2o65b2o5b2o40b3o$45bo$20b2o21b3o$20b2o20bo$
2b4o36b2o63b3o$2o4b2o100bo$2o5bo15b2o5bobo73b3o$2b2obobo15b2o5b2o$7bo
23bo$3bo3bo$3bo4bo129b2o$5b3o3bo126b2o$5b2o4bo$11b2o16bo111b2o$13bo15b
obo5b2o3b2o97b2o$13b3o13b2o8b3o$38bo3bo65b2o$39bobo66b2o28b2o$16bo23bo
89bo7b2o$15bob5o83b2o21b3o$14b2o5bo83b2o20bo$14b2o3bo2bo104b2o$22bo17b
2o44b4o$16b2obo2bo17b2o42b2o4b2o16b2o5bobo$19bo2bo61b2o5bo16b2o5b2o$
20b2o64b2obobo24bo$20b2o5bo4bo58bo$25b2ob4ob2o52bo3bo$27bo4bo54bo4bo$
89b3o3bo$89b2o4bo18bo$95b2o17bobo5b2o3b2o$97bo16b2o8b3o$97b3o23bo3bo$
124bobo$125bo$100bo$99bob5o$98b2o5bo$98b2o3bo2bo18b2o$106bo18b2o$100b
2obo2bo$103bo2bo$104b2o6bo4bo$104b2o4b2ob4ob2o$112bo4bo!
On the right is the standard Sawtooth 177, at size 74*60 (63 with pd sparks). Its minimum population is now achieved in gens 28*(121^n-1).

On the left is the compactified variant, at size 57*63 (60*63 with pd sparks). Its minimum population is 189(+10 at odd n), at gens 23*(121^n-1). The odd multiplier means that the minimum population only occurs every other cycle.

EDIT: Forgot to count the pentadecathlon sparks. Though the twin bees sparks weren't counted in Sawtooth 213 or 201.
"What's purple and commutes?
The Evanston Express."

User avatar
dvgrn
Moderator
Posts: 10612
Joined: May 17th, 2009, 11:00 pm
Location: Madison, WI
Contact:

Re: Smaller sawtooth

Post by dvgrn » October 31st, 2015, 10:40 am

thunk wrote:EDIT: Forgot to count the pentadecathlon sparks. Though the twin bees sparks weren't counted in Sawtooth 213 or 201.
Yup, the bounding box of Sawtooth 201 would be larger (79x59) if sparks were included in the same way that they are traditionally included in glider gun bounding boxes.

Luckily, sawteeth are not glider guns -- their bounding boxes are always increasing in any case. And clearly there's something of an established tradition to measure the bounding box without any transitory sparks, unless they're actually present in the T=0 generation. So I've added the new Sawtooth 189 record as 63x57 in the LifeWiki and left all the old bounding-box measurements unchanged.

There's a potentially difficult case that comes up here, where a sawtooth with a smaller bounding box can be constructed if the minimum-population "canonical phase" occurs at T>0. Luckily I think that's not an issue for Sawtooth 189, since the spaceship is already as close as it can safely get to the pentadecathlon.

thunk
Posts: 170
Joined: October 3rd, 2015, 8:50 pm
Location: Central USA

Re: Smaller sawtooth

Post by thunk » October 31st, 2015, 11:45 am

dvgrn wrote:Luckily, sawteeth are not glider guns -- their bounding boxes are always increasing in any case. And clearly there's something of an established tradition to measure the bounding box without any transitory sparks, unless they're actually present in the T=0 generation. So I've added the new Sawtooth 189 record as 67x57 in the LifeWiki and left all the old bounding-box measurements unchanged.
Ah, okay.

Anyway, I forgot to mention: Since we don't need the buckaroo to be synchronized with the pd, the pd can be moved one cell further in, resulting in a bounding box of 62*56.

Code: Select all

x = 62, y = 56, rule = B3/S23
53b2obo$53bob2o2$11b2o$11b2o3$14b2o28b2o7b2o5b2o$4b2o8b2o37b2o5b2o$5bo
37bobo$4bo6b2o28bo14b2o$4b2o5b2o24b2obob2o12b2o$38bob2o$39bo2$33b2o$8b
2o23b2o$8bobo$8bo20b2o5b2o$29b2o5b2o4$21b3o$21bo2bo$21bob2o4$20b2o12bo
9bo$20bo2bo10bobo5b3o$2b4o15b3o10b2o5bo$2o4b2o33b2o$2o5bo$2b2obobo14b
2o$7bo14b2o$3bo3bo26bo4bo$3bo4bo23b2o5bo$5b3o3bo21b2o3bobo$5b2o4bo25b
2ob2o$11b2o23bo5bo$13bo25bo$13b3o20b2o3b2o3$16bo21bo$15bob5o16bo$14b2o
5bo15bo$14b2o3bo2bo$22bo$16b2obo2bo16b2o$19bo2bo16b2o$20b2o$20b2o5bo4b
o$25b2ob4ob2o$27bo4bo! 
Unfortunately, the minimum population is increased still further to 195, achieved in gens 38*(121^n-1) (every cycle now).
"What's purple and commutes?
The Evanston Express."

User avatar
dvgrn
Moderator
Posts: 10612
Joined: May 17th, 2009, 11:00 pm
Location: Madison, WI
Contact:

Re: Smaller sawtooth

Post by dvgrn » October 31st, 2015, 12:24 pm

thunk wrote:Since we don't need the buckaroo to be synchronized with the pd, the pd can be moved one cell further in, resulting in a bounding box of 62*56...
Unfortunately, the minimum population is increased still further to 195, achieved in gens 38*(121^n-1) (every cycle now).
That doesn't seem particularly unfortunate. If you're optimizing the bounding box, the minimum population doesn't matter, and vice versa -- unless you can hold one metric constant and reduce the other, anyway.

I've done the necessary quick fixes to the LifeWiki, and am now trying to pretend that the almost-smallest almost-lowest-population Sawtooth 189 was never mentioned. Someone can create a page for Sawtooth 195 if they want, but I'm pretty happy just calling it a variant of Sawtooth 177 and providing an external link to the pattern, here in this thread.

There isn't a separate section in Sawtooth 177 showing Sawtooth 195 (yet -- that would be another option). So for now I just added SneakyWikiLinks: "[[Sawtooth 177|Sawtooth 195]]".

User avatar
calcyman
Moderator
Posts: 2932
Joined: June 1st, 2009, 4:32 pm

Re: Smaller sawtooth

Post by calcyman » October 31st, 2015, 1:25 pm

dvgrn wrote:Luckily, sawteeth are not glider guns -- their bounding boxes are always increasing in any case. And clearly there's something of an established tradition to measure the bounding box without any transitory sparks, unless they're actually present in the T=0 generation.
Where do you stand on the following variant of Sawtooth 177? It attains its minimum repeating population of 177 in generation 0, so it's canonical, and is 74-by-58 instead of 74-by-60. But it feels incredibly disingenuous in a way that is impossible to capture mathematically:

Code: Select all

#C Canonical Sawtooth 177 with reduced bounding box:
x = 74, y = 58, rule = B3/S23
64b4o2$66b2o$62b2o$62b2o8b2o$13b2o57bo$13b2o50b2o6bo$65b2o5b2o2$16b2o$
6b2o8b2o$7bo46b3o$6bo6b2o40bo$6b2o5b2o40b3o4$23b3o$24bo$22b3o4$54b2o$
54b2o2$57b2o$57b2o2$24b2o$24b2o28b2o$46bo7b2o$21b2o21b3o$21b2o20bo$43b
2o$2b4o$2o4b2o16b2o5bobo$2o5bo16b2o5b2o$2b2obobo24bo$7bo$3bo3bo$3bo4bo
$5b3o3bo$5b2o4bo18bo$11b2o17bobo5b2o3b2o$13bo16b2o8b3o$13b3o23bo3bo$
40bobo$41bo$16bo$15bob5o$14b2o5bo$14b2o3bo2bo18b2o$22bo18b2o$16b2obo2b
o$19bo2bo$20b2o$20b2o4b10o!
If anyone can see any reason why this shouldn't be considered the canonical Sawtooth 177, speak now...
What do you do with ill crystallographers? Take them to the mono-clinic!

User avatar
dvgrn
Moderator
Posts: 10612
Joined: May 17th, 2009, 11:00 pm
Location: Madison, WI
Contact:

Re: Smaller sawtooth

Post by dvgrn » October 31st, 2015, 1:59 pm

calcyman wrote:Where do you stand on the following variant of Sawtooth 177? It attains its minimum repeating population of 177 in generation 0, so it's canonical, and is 74-by-58 instead of 74-by-60. But it feels incredibly disingenuous in a way that is impossible to capture mathematically...
Bah, humbug. I mean... I was definitely thinking about tricks like this this morning, but figured I wouldn't speak up and give anyone any horrible ideas.
calcyman wrote:If anyone can see any reason why this shouldn't be considered the canonical Sawtooth 177, speak now...
I think every sawtooth up to now has returned to its exact initial configuration every time it has returned to its minimum population. This unfortunately isn't required by the definition of a sawtooth... but I think we could safely say that "canonical" implies "the phase of the sawtooth that you can compare to every future minimum phase, and see that it's the same except for the position of the spaceship."

Also the T=1 population doesn't match the pattern that all future sawtooth cycles will follow -- there are a couple of extra cells still present in the pre-pentadecathlon.

... If the T=1 population-graph argument doesn't hold, then another possible reason that this isn't really the canonical Sawtooth 177, is that someone might well be able to make an even more evil and disingenuous version that cuts off another few rows or columns. There are at least a dozen cells available that could be harmlessly subtracted from their starting location and added anywhere they're needed, to make pre-blocks or a pre-snake.

I'm not sure that this can actually be done, and I'm trying hard not to think about it. It's a valid optimization problem, I suppose, not so different from shrinking a spacefiller down into its smallest possible bounding box, or curling up the spiral-growth recipe gliders so they fit into the initial diamond shape... but I think it's worth defending the definition of "canonical" so that it continues to mean the 74x60 Sawtooth 177.

User avatar
dvgrn
Moderator
Posts: 10612
Joined: May 17th, 2009, 11:00 pm
Location: Madison, WI
Contact:

Re: Smaller sawtooth

Post by dvgrn » October 31st, 2015, 2:11 pm

dvgrn wrote:Also the T=1 population doesn't match the pattern that all future sawtooth cycles will follow -- there are a couple of extra cells still present in the pre-pentadecathlon.
Unfortunately the population graph can be made to match exactly, with hardly any trouble at all:

Code: Select all

x = 74, y = 58, rule = B3/S23
64b4o2$66b2o$62b2o$62b2o8b2o$13b2o57bo$13b2o50b2o6bo$65b2o5b2o2$16b2o$
6b2o8b2o$7bo46b3o$6bo6b2o40bo$6b2o5b2o40b3o4$23b3o$24bo$22b3o4$54b2o$
54b2o2$57b2o$57bo$59bo$24b2o$24b2o28b2o$46bo7bo$21b2o21b3o9bo$21b2o20b
o$43b2o$2b4o$2o4b2o16b2o5bobo$2o5bo16b2o5b2o$2b2obobo24bo$7bo$3bo3bo$
3bo4bo$5b3o3bo$5b2o4bo18bo$11b2o17bobo5b2o3b2o$13bo16b2o8b3o$13b3o23bo
3bo$40bobo$41bo$16bo$15bob5o$14b2o5bo$14b2o3bo2bo18b2o$22bo18b2o$16b2o
bo2bo$19bo2bo$20b2o$20b2o4b10o!
So I'm afraid the definition of "canonical" is going to have to carry all the weight of defending sawtooth initial bounding boxes against optimizations from the Dark Side. Anyone else have a better defense?

chris_c
Posts: 966
Joined: June 28th, 2014, 7:15 am

Re: Smaller sawtooth

Post by chris_c » October 31st, 2015, 2:37 pm

dvgrn wrote: So I'm afraid the definition of "canonical" is going to have to carry all the weight of defending sawtooth initial bounding boxes against optimizations from the Dark Side. Anyone else have a better defense?
Suppose PatternA and PatternB are identical after N ticks. Let S be the set of cells that become periodic in both patterns. We can say that PatternA <= PatternB if each cell in S becomes periodic in PatternA not later than the time that it becomes periodic in PatternB.

That should cut out the tricks mentioned so far. Maybe there are other problems with the definition but I'm trying not to look too hard for them.

User avatar
dvgrn
Moderator
Posts: 10612
Joined: May 17th, 2009, 11:00 pm
Location: Madison, WI
Contact:

Re: Smaller sawtooth

Post by dvgrn » October 31st, 2015, 8:56 pm

calcyman wrote:
thunk wrote:
Scorbie wrote:If I got it right, all the components are in its minimal phase, so there won't be any more improvements with trivial (i.e. without change of mechanism) collorary sniping, right??
I think so too...
Yes, I'm pretty certain it is optimal based on present technology.
dvgrn wrote:Luckily I think that's not an issue for Sawtooth 189, since the spaceship is already as close as it can safely get to the pentadecathlon.
Um... for Sawtooth 177, however, it's perfectly possible to move the spaceship one diagonal step closer, cutting the bounding box by one column. No Dark Side trickery, really, just Yet Another Corollary-Snipe:

Code: Select all

#C population 177 at T=0, 336720, 4930254240, ... 23*(121^2N-1)...
x = 73, y = 60, rule = B3/S23
64b2o$64b2o3$61b2o$61b2o8b2o$12b2o57bo$12b2o50b2o6bo$64b2o5b2o2$15b2o$
5b2o8b2o$6bo46b3o$5bo6b2o40bo$5b2o5b2o40b3o4$22b3o$23bo$21b3o4$53b2o$
53b2o2$56b2o$56b2o2$23b2o$23b2o28b2o$45bo7b2o$20b2o21b3o$20b2o20bo$2b
4o36b2o$2o4b2o$2o5bo15b2o5bobo$2b2obobo15b2o5b2o$7bo23bo$3bo3bo$3bo4bo
$5b3o3bo$5b2o4bo$11b2o16bo$13bo15bobo5b2o3b2o$13b3o13b2o8b3o$38bo3bo$
39bobo$16bo23bo$15bob5o$14b2o5bo$14b2o3bo2bo$22bo17b2o$16b2obo2bo17b2o
$19bo2bo$20b2o$20b2o5bo4bo$25b2ob4ob2o$27bo4bo!
The multiplier goes down from 28 to 23, the same odd multiplier as the compactified Sawtooth 189, so the sawtooth only reaches its minimum every other cycle. At 23*(121^N-1) for odd N, the minimum is 187 due to the two boat-bits, so the expansion factor goes up to 121^2 = 14641.

Post Reply