1. Forum
  2. Usenet
  3. REC.GAMES.COREWAR

  • Epic Fail

    From Stephen Gunnell@21:1/5 to All on Tue Sep 13 18:44:54 2022
    For nearly 7 years now I have been running my 94 standard workspace with p-warriors crippled. Sometime back before 2015 I split my 94 workspace into p-space and non-p-space versions. Soon afterwards when Windows 10 tried to force its was onto my desktop
    I switched back to Linux. Then I was having problems with p-space warriors contaminating my 94nop workspace I set S=1 in my pmars parameter file for 94nop to weed them out. Unfortunately after I migrated my workspaces from Windows to Linux I ran my usual
    script to convert duplicate copies of files to hard links I linked the 94standard parameters with the 94nop parameters. Initially there was no change because the results of historic battles were cached in a database but it became increasingly difficult
    to get new p-space warriors to score well and new 94nop warriors score increasingly well against old p-space warriors.
    Finally I spotted some odd behavior in a p^3 switcher which caused me to waste a bunch of time looking for a bug in the generator which eventually led me to suspect a problem with the amount of p-space being allocated by pmars. I went to the parameter
    file to force a value to check my suspicions and found my 94nop parameter laughing at me. Sheesh!

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Skybuck Flying@21:1/5 to stephen.g...@gmail.com on Thu Sep 22 10:37:48 2022
    On Wednesday, September 14, 2022 at 3:44:55 AM UTC+2, stephen.g...@gmail.com wrote:
    For nearly 7 years now I have been running my 94 standard workspace with p-warriors crippled. Sometime back before 2015 I split my 94 workspace into p-space and non-p-space versions. Soon afterwards when Windows 10 tried to force its was onto my
    desktop I switched back to Linux. Then I was having problems with p-space warriors contaminating my 94nop workspace I set S=1 in my pmars parameter file for 94nop to weed them out. Unfortunately after I migrated my workspaces from Windows to Linux I ran
    my usual script to convert duplicate copies of files to hard links I linked the 94standard parameters with the 94nop parameters. Initially there was no change because the results of historic battles were cached in a database but it became increasingly
    difficult to get new p-space warriors to score well and new 94nop warriors score increasingly well against old p-space warriors.
    Finally I spotted some odd behavior in a p^3 switcher which caused me to waste a bunch of time looking for a bug in the generator which eventually led me to suspect a problem with the amount of p-space being allocated by pmars. I went to the parameter
    file to force a value to check my suspicions and found my 94nop parameter laughing at me. Sheesh!

    I am glad you finally realized that you are a BIG FAT FAIL ! LOL.

    Something I already detected during our Corona and DEAD BABIES discussion ! LOL.

    :P******************* =DDDD

    Bye,
    Skybuck ! ;) =D

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From inversed@21:1/5 to All on Sun Oct 16 02:04:09 2022
    I won't be outdone when it comes to epic fails.

    The biggest one is losing the Valentine tournament 2015. I'm pretty sure I would've won had I sent just one entry (Infiltrator) instead of two. I knew there would be no counters to it. I predicted Steve's strategy. But I thought there would be no harm in
    sending a second entry. WRONG. The other entry (Monolith) advanced, which cost me the tournament.

    Something similar happened during the recent 2022 CGM finals when the wrong entry advanced. This was less of a fail: the outcome was difficult to predict and Roy would've won anyway thanks to a better strategy (would've been a much closer fight though).
    The real mistake was spending most of the time on the supernano round (it interested me the most) and preparing for the finals hastily.

    The latest fail is not fully understanding my own q^i quickscan design. This year I had a "what was I thinking!?" moment when I realized that the additional decode stage is unnecessary. At least this one is fixable.

    --
    inversed

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Stephen Gunnell@21:1/5 to inversed on Sat Oct 29 15:49:00 2022
    On Sunday, 16 October 2022 at 5:04:10 pm UTC+8, inversed wrote:
    I won't be outdone when it comes to epic fails.

    Heh, I know those "what was I thinking?" moments all to well. Since I retired I have re-visited some never quite good enough code and have done some forehead slapping myself.

    You can only fail spectacularly if you are actually pushing the envelope.

    Cheers,
    Steve

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From John Metcalf@21:1/5 to Stephen Gunnell on Sat Nov 5 17:06:24 2022
    On Sat, 29 Oct 2022 15:49:00 -0700, Stephen Gunnell wrote:

    On Sunday, 16 October 2022 at 5:04:10 pm UTC+8, inversed wrote:
    I won't be outdone when it comes to epic fails.

    Heh, I know those "what was I thinking?" moments all to well. Since I
    retired I have re-visited some never quite good enough code and have
    done some forehead slapping myself.

    You can only fail spectacularly if you are actually pushing the
    envelope.

    Cheers,
    Steve

    I'd say your successes far outweigh any fails :-) It happens to everyone
    at some point. I wasted months optimizing for the normal hill using a
    special tournament version of pMARS which filled empty core with DAT 1,1!

    By the way, you can switch off p-space in config.h

    John

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)