www.ClassicTW.com :: View topic - Need someone to Test proZTM3.1 / ZTM Test Results

www.ClassicTW.com https://mail.black-squirrel.com/

Need someone to Test proZTM3.1 / ZTM Test Results https://mail.black-squirrel.com/viewtopic.php?f=15&t=21205	Page 1 of 1

Author:

Promethius [ Mon Nov 24, 2008 12:12 pm ]

Post subject:

Need someone to Test proZTM3.1 / ZTM Test Results

I need someone to test a script that can check the accuracty of proZTM3.1. I ran tests today on a 5k bang and think I just got lucky on the results or maybe it was the fact I was getting tired. To check the accuracy, you will need the warps from a 5k edit from an observer account, and a small code snippet I will provide that checks a scripts output accuracy. Yes, proZTM does provide this, but I am running the snippet anyway against all ZTM's I am testing.

The results of my testing is below. Take my results on proZTM3.1 with a very large grain of salt and also the time elapsed - was having "fun" doing the time calcs by hand.

Observer Account

Observer Account 5k Edit
 959 sectors with  1 warps out.
1634 sectors with  2 warps out.
1274 sectors with  3 warps out.
 678 sectors with  4 warps out.
 299 sectors with  5 warps out.
 156 sectors with  6 warps out.

Total warps found:  13192

ztmSupG started 21:56:21
Pass      Complete    Warps   Elapsed
 1        22:11:51    10559   0:15:30
 2(v1-4)  23:01:39    12931   1:05:18  
 3        23:05:15    13176   1:08:54

 961 sectors with  1 warps out.
1633 sectors with  2 warps out.
1274 sectors with  3 warps out.
 683 sectors with  4 warps out.
 297 sectors with  5 warps out.
 152 sectors with  6 warps out.

Total warps found:  13178  99.89%

SWATH 00:50:00
Pass      Complete    Warps   Elapsed
 1 "var"  01:31:59    12724   0:41:59        
 2 "fold" 02:16:43    13125   1:25:43 (1min deduct)   
 3  1-w   02:18:54    13179   1:27:54                             

 968 sectors with  1 warps out.
1629 sectors with  2 warps out.
1270 sectors with  3 warps out.
 678 sectors with  4 warps out.
 299 sectors with  5 warps out.
 156 sectors with  6 warps out.

Total warps found:  13179  99.90%

proZTM3.1 started 04:45:38             
Pass      Complete    Warps   Elapsed
 1        05:01:17    10573   0:15:39
 2        05:15:15    12660   0:29:37
 3        05:59:53    13177   1:14:15
 4 & 5    06:07:59    13183   1:22:37

 959 sectors with  1 warps out.
1634 sectors with  2 warps out.
1274 sectors with  3 warps out.
 678 sectors with  4 warps out.
 299 sectors with  5 warps out.
 156 sectors with  6 warps out.

Total warps found:  13192    100% ????

Author:	Singularity [ Mon Nov 24, 2008 12:56 pm ]
Post subject:	Re: Need someone to Test proZTM3.1 / ZTM Test Results
The trick w/ most ZTM algos is complex bubble and loop structures. Take a look at CK's ztm, that's the way I modeled mine. It has different passes, a 1 pass can be used for basic gridding, a 2 pass can be used to clarify DEs, etc to the final 6th pass. Quality depending on the pass. Even on the highest quality tho you'll have some problems w/ complex bubbles or loops. To get that tho you'll need to bang a 20k game w/ big gold bubbles, and it'll be psuedo-random determined by the game seed.

Author:	Promethius [ Mon Nov 24, 2008 2:27 pm ]
Post subject:	Re: Need someone to Test proZTM3.1 / ZTM Test Results
Singularity wrote: The trick w/ most ZTM algos is complex bubble and loop structures. Take a look at CK's ztm, that's the way I modeled mine. It has different passes, a 1 pass can be used for basic gridding, a 2 pass can be used to clarify DEs, etc to the final 6th pass. Quality depending on the pass. Even on the highest quality tho you'll have some problems w/ complex bubbles or loops. To get that tho you'll need to bang a 20k game w/ big gold bubbles, and it'll be psuedo-random determined by the game seed. I ran CK's ztm, but all I have is the .cts and it ran in computer mode with a lot of "skipping xx" echos in the screen. I know he is just letting people know what it is doing, but irritating to me. I have put RiverCitytw.dlinkddns.com 2002 back online, and banged 3 games if anyone wants to test ZTMs. Games N 5k, O 10k, P 20k (Sub Zero) are available and the warp count from the observer account is shown for these three games on the game menu. And yes, I will prob be in the games since the server is not going active except for the testing.

Author:	Singularity [ Mon Nov 24, 2008 3:13 pm ]
Post subject:	Re: Need someone to Test proZTM3.1 / ZTM Test Results
Quote: I ran CK's ztm, but all I have is the .cts and it ran in computer mode with a lot of "skipping xx" echos in the screen. I know he is just letting people know what it is doing, but irritating to me. Nod, all I'm saying is that the method is a good benchmark and design method. Having multiple pass options means you can trade speed for accuracy during time-critical areas of the game. When determining accuracy then you need to have some idea as to the "pass quality" of the benchmark. It'd be pointless to compare a 2 pass CK to a full pass SupG, for instance. Still, I'd recommend testing against a ck 6 pass in the comparison. Anyway, a full multi-pass ztm will be pretty accurate. Minus complex structures (like bubbles) where setting an avoid blocks multiple paths out. You'll find those are missed regardless, unless you make a specific point of searching for them. Those structures will only exist in a 20k uni w/ large gold bubbles enabled.

Author:	Promethius [ Mon Nov 24, 2008 3:38 pm ]
Post subject:	Re: Need someone to Test proZTM3.1 / ZTM Test Results
Have to agree on the multi-pass options being better. With SupG's ztm, I ran with default settings as I believe that is what most users do with a script. The SWATH setting was also the default. I started testing on 5k edits due to the time required for a 20k ZTM. I figured if I could get the basic 5k down, then I could see what it would do on the more complicated edits. The 5k tests basically provide a benchmark to work off of - if the 5k isn't dead on, then the 20k will really be screwy. I did find a major issue this way with the proZTM scripts and that is somewhere along the line I messed up one the the passes, it ran, but didn't do much good. I have fixed that and a couple of other issues related to running split ZTMs for corps - just slowed it way down from where it should have been in that mode. As for configuration, I am making the 7-warps in window an option and also the total warps found during the pass. I will still have it show the total warps found after the pass. With the adjustment to the burst mode, the total warps found during the pass may not slow it down very much. One thing about showing the total warps found is that the user can determine if they think they have enough warps found for whatever activity they are doing, and then they can resume the ZTM later.

Author:	Promethius [ Mon Nov 24, 2008 8:59 pm ]
Post subject:	Re: Need someone to Test proZTM3.1 / ZTM Test Results
Well tested in a 20k universe - Sub Zero edit - and I didn't hit 100% accuracy... 20k Testing Observer Account = 52595 Warps proZTM3.1 started 13:35:23 Pass Complete Warps Elapsed Accuracy 1 14:42:03 42831 1:07:39 81.4% 2 15:44:27 49930 2:09:04 94.9% 3 19:07:15 52521 5:31:52 99.8% 4 19:43:13 52542 6:07:42 99.989% 5. 19:49:20 52589 6:14:49 99.998% I haven't checked the universe data to see what the 6 warps missed were.

Author:	LoneStar [ Tue Nov 25, 2008 7:57 am ]
Post subject:	Re: Need someone to Test proZTM3.1 / ZTM Test Results
Promethius wrote: Well tested in a 20k universe - Sub Zero edit - and I didn't hit 100% accuracy... 20k Testing Observer Account = 52595 Warps proZTM3.1 started 13:35:23 Pass Complete Warps Elapsed Accuracy 1 14:42:03 42831 1:07:39 81.4% 2 15:44:27 49930 2:09:04 94.9% 3 19:07:15 52521 5:31:52 99.8% 4 19:43:13 52542 6:07:42 99.989% 5. 19:49:20 52589 6:14:49 99.998% I haven't checked the universe data to see what the 6 warps missed were. Have you considered keeping track of course plots, in which there were no avoids and a course could not be made --the path exceeded the game maximum. While writing a Dock finder script for the HHT, I ran into this problem of paths exceeding 45 hops and got alot of "Error - No route within ..." msg's and managed to circumvent it by making alternative plots: A -> Z nope, B -> Z nope, ... -> Z Bingo!

Author:	Promethius [ Tue Nov 25, 2008 1:52 pm ]
Post subject:	Re: Need someone to Test proZTM3.1 / ZTM Test Results
LoneStar wrote: Promethius wrote: Well tested in a 20k universe - Sub Zero edit - and I didn't hit 100% accuracy... 20k Testing Observer Account = 52595 Warps proZTM3.1 started 13:35:23 Pass Complete Warps Elapsed Accuracy 1 14:42:03 42831 1:07:39 81.4% 2 15:44:27 49930 2:09:04 94.9% 3 19:07:15 52521 5:31:52 99.8% 4 19:43:13 52542 6:07:42 99.989% 5. 19:49:20 52589 6:14:49 99.998% I haven't checked the universe data to see what the 6 warps missed were. Have you considered keeping track of course plots, in which there were no avoids and a course could not be made --the path exceeded the game maximum. While writing a Dock finder script for the HHT, I ran into this problem of paths exceeding 45 hops and got alot of "Error - No route within ..." msg's and managed to circumvent it by making alternative plots: A -> Z nope, B -> Z nope, ... -> Z Bingo! I've ran into that type of game before - broke every public ZTM I tried. The sysOp banged with a max of xx warps then changed the max to xx + 10 warps or something of that sort. I believe the evil op was named Vader. The "broken" universe caused all kinds of problems with movement if I remember right - was a long time ago. The setup of exceeding the maximum game bang plots is designed to slow attackers down, and just create general havoc I believe, but what is does do is cause a non-scripter major problems as it did me at the time. As you noted, there are ways around that issue - takes a bit longer, and scripts can be adapted for the attack once the ZTM is complete - that is if a player can adapt. So, for those that think they've got mad skillz because of public scripts and can't script, what'cha gonna do now? Well, if you are a player and not just a gamer, you learn a bit about scripting - it isn't black magic. If I can write a script - even one that sends you to dock instead of terra to grab colos - anyone can.

Author:	Singularity [ Tue Nov 25, 2008 2:53 pm ]
Post subject:	Re: Need someone to Test proZTM3.1 / ZTM Test Results
Nod. Just build a ztm patcher, look for 0 warp in/out sectors and plot in and out to terra, dock, and a few random sectors if needed. It's a simple script to build.

Author:	V'Ger [ Tue Nov 25, 2008 3:10 pm ]
Post subject:	Re: Need someone to Test proZTM3.1 / ZTM Test Results
Promethius wrote: If I can write a script - even one that sends you to dock instead of terra to grab colos - anyone can. Now what would really be a good script is instead of just sending you to dock to get collies, it actually got them from there

Page 1 of 1	All times are UTC - 5 hours
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group http://www.phpbb.com/