Link to Null Skip detection attack spreadsheet by smokie and Jarlve Spring 2018:
https://docs.google.com/spreadsheets/d/ … sp=sharing

This is on the first post of this thread, about half way down, easy to find. Tonight I am going to start punching out some of the easy ones. Will probably get a good book and / or work on home improvement while the program is running over upcoming days / weeks / months etc. Thanks.

Link to Null Skip detection attack spreadsheet by smokie and Jarlve Spring 2018:
https://docs.google.com/spreadsheets/d/ … sp=sharing

This is on the first post of this thread, about half way down, easy to find. Tonight I am going to start punching out some of the easy ones. Will probably get a good book and / or work on home improvement while the program is running over upcoming days / weeks / months etc. Thanks.

It seems to be working. Yes, this may take months.

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.

I did. Are you okay with it? I like a healthy amount of HC iterations for these high nulls & skips counts. If the hill climber still can be improved allot then these numbers could be brought down but it may also be that these numbers will need to come up even further. We will see in due time.

I am very o.k. with it. Working on a simple one trial 25 now. It won’t solve, but I am going to save my outputs by trial number, interested in seeing if specific positions keep showing up in the results over and over again.

I am very o.k. with it. Working on a simple one trial 25 now. It won’t solve, but I am going to save my outputs by trial number, interested in seeing if specific positions keep showing up in the results over and over again.

I see you added the first entry. Should we not include the full AZdecrypt solve log so that we also know the positions of the nulls & skips?

Full AZdecrypt solve log:

Score: 24083.40 Ioc: 0.06647
Ngrams: 1800 PC-cycles: 6914

Key: Period(20) Nulls(43,130,131,194,247), Skips(72,169,231)

YESIWASPROSTITUTE
ITWASNTMYFINESTHO
URANDITSNOTHINGTH
ATIMATALLASHAMEDO
FWEALLUSEDTOBESOM
ETHINGTHATWEMAYOR
MAYNOTBEPROUDOFIT
SALSONOTLOSTONMET
HEOBVIOUSRESEMBLA
NCEOFMINEANDOFFIC
IALLCAFEESRELATIO
NSHIPTOTHATOFRICH
ARDGEREAWAVEOFTRA
SHTHEEMRSMCAFEEIS
HAPPENINGBECAUSEO
FHATREDTOWARDSMEY
ESSHEWASAPROSTITU
TEGETOVERITPICTUR
ESPAINTATHOUSANDW
ORDSBELOWTHEDAY

I would have liked to and tried to make the spreadsheet do that. When I paste the full solve log into the spreadsheet, the result goes into a lot of different rows. Go the paste utility on the next tab at the bottom. Paste the solve into the green shaded cell and it will concatenate into one cell above. Then ctrl C to copy that, and go to the trial log spreadsheet and ctrl shift V to paste the value into the solve column.

I was thinking about making a separate column for nulls skips part of the solve log, but was just home from work and wasn’t sure what you would want. I saved my trial output for now. Be my guest to modify. Trying to figure out how to get my windows 10 explorer to show the preview field / pane so I can scroll through a lot of them to look for consistency, but can’t so far. I am on trial 27 now.

Okay, you could also paste the solve log in the AZdecrypt input window, add one space after the score, ngrams and key lines and go to Format and click on Remove line breaks and you get it all on one line for the spreadsheet:

Score: 24083.40 Ioc: 0.06647 Ngrams: 1800 PC-cycles: 6914 Key: Period(20) Nulls(43,130,131,194,247), Skips(72,169,231) YESIWASPROSTITUTEITWASNTMYFINESTHOURANDITSNOTHINGTHATIMATALLASHAMEDOFWEALLUSEDTOBESOMETHINGTHATWEMAYORMAYNOTBEPROUDOFITSALSONOTLOSTONMETHEOBVIOUSRESEMBLANCEOFMINEANDOFFICIALLCAFEESRELATIONSHIPTOTHATOFRICHARDGEREAWAVEOFTRASHTHEEMRSMCAFEEISHAPPENINGBECAUSEOFHATREDTOWARDSMEYESSHEWASAPROSTITUTEGETOVERITPICTURESPAINTATHOUSANDWORDSBELOWTHEDAY

Made my first entry. This will be fun.

The entry mirrors my finding in viewtopic.php?f=81&t=3206

Skip at position 200, right on a pivot.

Looking good. I like the little notes. Working on trial 28 now, but all work day, which is somewhat of a waste. I will get my schedule and trials more appropriate with each other soon.

Looking good. I like the little notes. Working on trial 28 now, but all work day, which is somewhat of a waste. I will get my schedule and trials more appropriate with each other soon.

It will be okay. I may enter some large number of restarts at times depending if the computer was left running while I went to sleep or work, or both.

Here are again the solve rates for jarlve_p20_5nulls_3skips with added entries:

10k HC iterations: 1.93%
20k HC iterations: 4.20%
40k HC iterations: 9.38%
80k HC iterations: 16.66%

Going from 40k to 80k the solve rate did not double as with previous jumps which indicates that starting around 40k, restarts become more efficient for the difficulty of this cipher. What I want to test now is 1000000 substitution iterations instead of 500000 (double) and see how much it affects the solve rates. And eventually also test 6-grams solve rates.

@smokie, I have released AZdecrypt 1.11: viewtopic.php?f=81&t=3198

There is an added option for the nulls & skips solver called bias. If left at 50, the solver will select a null/skip with 50%/50% chance. If at 25, it is 25%/75% chance. If at 0 the solver can only use skips and if at 100 the solver can only use nulls. At 50 this release works exactly the same as version 4.3.18.

O.k.

I am going up to 6 nulls and skips right now, should I go higher?

O.k.

I am going up to 6 nulls and skips right now, should I go higher?

Go up to 8. We still need to confirm a solve on 8+. That may require further optimization of the hill climber first.

I checked all 1,953 combinations of two symbols in the 340, deleting them as if they were nulls. Only 7 ( .36% ) of the combinations resulted in a net increase in P19, and the one that caused the largest increase was the W – theta combination, the two regional bias symbols.

I checked all 1,953 combinations of two symbols in the 340, deleting them as if they were nulls. Only 7 ( .36% ) of the combinations resulted in a net increase in P19, and the one that caused the largest increase was the W – theta combination, the two regional bias symbols.

That is probably because these symbols are near the start and end of the cipher. I have noted this behaviour before in nulls & skips ciphers. Say you have a transposition cipher with one null and run a hill climber on it, the best improvements will be found in the vicinity of the null. Now remove the null and run the hill climber again and it will find the best improvements near the start or end of the cipher. Why? These positions least affect the order after untransposition.