Nulls would be intentional filler, gibberish, or a shape or hidden additional message, etc. But skips would be a mistake. I think that there would probably be a lot more nulls than skips.
I do not want to make that assumption.
Nulls(43,118,137,194,246), Skips(70,173,231) partial solve
Nulls(43,130,131,195,247), Skips(72,169,231) perfect or near perfect solveI see that the positions for the partial solve are the same or really close to the positions for the perfect solve, except for the two side by side nulls. Those are a lot farther off, one by 12 positions and one by 6 positions.
Now there is a 10% chance to shift a null/skip randomly between 1 and half a period 20 column to the right or left. A period 20 column is 17 long so changing it to between 1 and a period 20 column may better the hill climber’s performance on my cipher since it will be able to find the 12 positions off easier.
Here is the link to the spreadsheet. I made it so anyone can view, but only you and I can edit. I used your yahoo address, so not sure if that will work. Interesting to know, otherwise p.m. me a gmail address.
It is looking good.
I have added "jarlve was here", can you see it?
If we want to add restarts to a test we just update the fields that have changed right?
It creates a local copy for me. So any changes are not shared. Hmmm. I think you need to copy paste a shareable link.
I assume you are logged into yahoo. The application or cloud space is named "Google Drive" so you may need a Gmail account. The application has a feature that copies a shareable link to the clipboard that is what I did.
If we want to add restarts to a test we just update the fields that have changed right?
If you want trials with 10000, 20000, and 40000 HC we can just insert a row I suppose.
I assume you are logged into yahoo. The application or cloud space is named "Google Drive" so you may need a Gmail account. The application has a feature that copies a shareable link to the clipboard that is what I did.
You need to click on Share in Google Spreadsheet itself with your document open and then enter my yahoo e-mail.
If we want to add restarts to a test we just update the fields that have changed right?
If you want trials with 10000, 20000, and 40000 HC we can just insert a row I suppose.
Yes and allow for even more in the future.
By the way, I could add a setting called nulls & skips bias. A value from 0 to 100. If at 50 the bias of selecting a null or skip is 50/50. If at 25 the bias is 25/75 (25% for a null). If at 75 the bias is 75/25. You get the idea, that would be very simple to add.
I assume you are logged into yahoo. The application or cloud space is named "Google Drive" so you may need a Gmail account. The application has a feature that copies a shareable link to the clipboard that is what I did.
You need to click on Share in Google Spreadsheet itself with your document open and then enter my yahoo e-mail.
Sorry about the confusion. This is what I did. If I select you it says you can edit, but when I try to use the snipping tool it won’t show that drop down menu. Now I will try what you said.
Still working on it. May take a little while to get this worked out.
If we want to add restarts to a test we just update the fields that have changed right?
If you want trials with 10000, 20000, and 40000 HC we can just insert a row I suppose.
We could avoid this by agreeing on a healthy amount of HC iterations up to a certain amount of nulls & skips. Say up to 6 nulls & skips 10000 HC iterations is enough. But starting from 7 move up to 40000 or similar. And from 9 even higher. I will motivate this reasoning later. But harder problems just do better with higher HC iterations versus the computer time invested so it makes sense.
I see it like this. Period 20 gives 20 columns, and to fix a null or skip you only need to know wether it is a null or a skip and in which column it is. That gives 20 times 2 (columns times null or skip) possibilities per suspected null/skip. The following table then gives the raw combinations/searchspace per amount of nulls/skips. Thank god for hill climbing.
1 null/skip: 40 possibilities 2 nulls/skips: 1600 possibilities 3 nulls/skips: 64000 possibilities 4 nulls/skips: 2560000 possibilities 5 nulls/skips: 102400000 possibilities 6 nulls/skips: 4096000000 possibilities 7 nulls/skips: 163840000000 possibilities 8 nulls/skips: 6553600000000 possibilities 9 nulls/skips: 262144000000000 possibilities 10 nulls/skips: 10485760000000000 possibilities
Let me ask a question. The null skip detection hill climber makes small, random changes to a group of null and skip positions, and then the substitution hill climber works on that for a while to determine if the random change in the null and skip position needs to be kept or not kept, right? Is that how it works? So, the program is getting really good partial solutions but the null or skip positions are off just by a few positions a lot of the times. Setting aside the recent conversation about fine tuning, would the program be more efficient at finding these near perfect positions if it make the small random changes to the null and skip positions by choosing from only say, odd or even positions?
I will be at a better connection later and work on getting you access to the spreadsheet.
Let me ask a question. The null skip detection hill climber makes small, random changes to a group of null and skip positions, and then the substitution hill climber works on that for a while to determine if the random change in the null and skip position needs to be kept or not kept, right? Is that how it works? So, the program is getting really good partial solutions but the null or skip positions are off just by a few positions a lot of the times. Setting aside the recent conversation about fine tuning, would the program be more efficient at finding these near perfect positions if it make the small random changes to the null and skip positions by choosing from only say, odd or even positions?
No it does not matter. In period 20 we have 20 columns. If then a random number between 1 and 340 is chosen it has a 5% chance to fall into any one column. Since any one column in period 20 takes up 17 positions. Change that to a random number between 1 and 340 using only even numbers gives the possibilities 2, 4, 6, 8, 10, etc… and the same 5% chance to fall into any one column because the positions of the period 20 column have also halved.
Try this link. I opened it in Google Sheets, and shared it with you.
We could avoid this by agreeing on a healthy amount of HC iterations up to a certain amount of nulls & skips. Say up to 6 nulls & skips 10000 HC iterations is enough. But starting from 7 move up to 40000 or similar. And from 9 even higher. I will motivate this reasoning later. But harder problems just do better with higher HC iterations versus the computer time invested so it makes sense.
I made changes to the spreadsheet to show what you are talking about. Feel free to change. I have never used a shared spreadsheet before so bear with me. We will figure it out.
