[1.20a] Strategic Pop Lock Analysis (i.e. Couch Potatoe [sp?] Factions Eating Bon Bons and Watching Springer Reruns)

from Elemental Forums

Purpose: Analyze strategic AI behavior to identify issues with factions pop locking.

Definition: Pop locking is defined as a faction sitting at max initial population of starting capital and not building improvements, units, etc. Not doing anything to expand their faction. Watching Oprah, Springer, or Judge Judy.

Scope: Cycling 36 turns of same map from the same turn 1 save, same factions, same starting locations, same AI difficulty level, and assessing whether a faction is pop locked at turn 36. Do this over 10 different games, record results, and analyze.

Scenario:

  • Large custom map built from "Ovia" - created an island so I can cycle turns, reduced starting positions to 10 and spread them out
  • Ridiculous across the board
  • Max players
  • Normal pacing
  • Normal sov
  • Playing as Pariden - thus not analyzed as part of this test
  • Used Ctrl-U cheat to see entire map
  • Used Ctrl-Z cheat to cycle turns

Data:

Results:

  • Gilden, Capitar, & Kraxis all pop locked twice each over 10 games - 20% chance of that occurring for those factions
  • None of the other 6 factions (Pariden unknown since ran test as player Pariden) ever experienced a pop lock
  • None of the ten games ever had more than 1 faction at a time pop lock
  • Overall 7% chance of a faction pop locking across all 10 games and 9 factions

Since I don't know how Brad has the AI coded I'm unsure why this is occurring. There are definitely 3 factions with issues for this particular map, on this particular difficulty, and their particular starting locations. Hayelp Brad or Derek! :cylon:

41,452 views 17 replies
Reply #1 from Elemental Forums Top

I mispelled potato. What a noob.

Reply #2 from Elemental Forums Top

lol I would think you can edit the title in the thread you know...

Elemental: Reforged
Reply #3 from Elemental Forums Top

Quoting EternalRequiem, reply 2
lol I would think you can edit the title in the thread you know...
End of EternalRequiem's quote

Double noob for not editing the title!

Reply #4 from Elemental Forums Top

you're your own worst enemy XD

Reply #5 from Elemental Forums Top

Removed - double post.

 

Reply #6 from Elemental Forums Top

Ok so as I watched reruns of Gilligan's Island I decided to run games 11-20 using the same baseline save as the first 10. I assigned each starting location a # to see if there is consistency across future tests. Below are the results:

  • Neither Kraxis or Capitar experienced further pop locks in 11-20. Their overall % dropped from 20% to 10%.
  • However, Gilden actually got worse with 3 pop locks. Total pop locks increased from 20% to 25%.

I think the next thing I'll do is create a brand new baseline turn 1 save with new starting locations and I'll make Pariden AI with Yithril as the player (since they experienced no pop locks). I'll track the starting location # to see if that has any bearing.

Reply #7 from Elemental Forums Top

Ok so here is my college thesis. I started a new game, same map but only changed the following: switched Pariden with Yithril (now testing Pariden and playing as Yithril) and the starting locations are different. Ran 10 games with new baseline save:

Baseline save - http://dl.dropbox.com/u/19969679/Elemental/119n/119n_POPLOCKTESTBASELINE2_1.EleSav

Here is the map I'm using. It's the map I normally use to analyze the strategic AI. I noted the location # on this map and identified the location of each faction across both rounds of tests. The player always starts on the island in the bottom right corner.

Ok so below are the results of the tests.

  • Round1 - the first 10 games from the initial analysis posts above this one
  • Round2 - the 10 games analyzed with the new baseline save

Summary:

  • By Faction:
    • Markinn from Gilden is one messed up dude - in round1 he was in location 3 and pop locked 20% of the time. In round 2, he was in location #1 and pop locked 50% of the time. Conversely, location 3 (Resoln) recorded ZERO pop locks in round2 and location 1 (Altar) recorded ZERO pop locks in round1. Thus, the location is not causing Markinn's nuttiness. Markin plain needs an intervention and some Braderly love.
    • Capitar improved in round 2. Was in a different location. He still needs to be looked at.
    • Kraxis definitely improved and recorded 0 pop locks - worth a looksie.
  • By Location:
    • Location 1 was perfect in round1 but horrible in round2 - I blame Gilden.
    • Location 2 was perfect in round1 but bad in round2 - Pariden problem? They were not tested in round1 because I was using them as player.
    • Location 6 was perfect in round1 but bad in round2 - I blame Capitar.
    • Location 8 experienced 2 pop locks on both rounds - problem with the location? Not sure why it has the resources.
    • Location 10 slightly improved in round2 - problem with the location? Not sure why it has the resources.

I double checked the map locations and all have fertile lands near the spawns - so it's not a resource problem.

Brad - if you need the saves of the results please let me know, I saved each game's turn 36. Thanks for looking into it. :cylon:

Reply #9 from Elemental Forums Top

Quoting EternalRequiem, reply 8
You the man
End of EternalRequiem's quote

Thanks. At this point Brad either loves me or hates me - but I guarantee he's not sitting on the fence, lol. :frogboy:

Reply #10 from Elemental Forums Top

1.19n issue. Speak of the devil, just started a brand new ridiculous game on random large vanilla map and Capitar died to monsters and ruffians on turn 10. I guarantee this is from not founding his capital - he seems to have a serious problem with this now and then. ARGH!

Zip (last 3 saves) - http://dl.dropbox.com/u/19969679/Elemental/119n/119n_10_CAPITARDEAD.zip

Reply #11 from Elemental Forums Top

I'm pretty damn sure the developers love you for this feed back, it helps them narrow down issues they can take care of, one of the major issues as this points out is... AI, wow... the AI is dumber then a door nob, thicker then a wall and just plain :dur:

Reply #12 from Elemental Forums Top

Ok so here is my doctorate research. I started a new game using the vanilla MEDIUM map Colaxi. I wanted to see if the previous map itself or size of the map affected results. I reduced the # of games from 10 to 5 but still did two rounds. In short all of the Kingdom factions are messed up with the exception of Altar. Kraxis needs a peek. Please note I cycled turns until 36, saved, and recorded results.

Here is the map and the locations noted:

Here are the baseline saves (turn 1 of each round):

Here are the results:

Summary:

  • By Faction:
    • Capitar batted a 1000 in the first game by being the only faction to pop lock AND pop lock every single game. Remarkable achievement of... underachievement! I've always thought Carrodus was an arrogant and lazy sob - round 1 proves it. Note his round1 debacle was at location 7 of which Altar sat in in round2 and NEVER had a pop lock. Also note he pop locked once in round 2 at location 1 - location 1 had zero pop locks in round 1 when Pariden had it.
    • I thought Markinn cleaned up his act but fell off the wagon in round2.
    • Pariden was the Biggest Loser of round2. Porcupine went from 0 in round1 at location 1 to 4 pop locks in round2 at location 3.
    • Tarth had a couple pop locks. That's ok, she's hawt.
    • Honestly, all of the Kingdom factions need to be reviewed with maybe the exception of Altar who has done fine in all testing thus far.
  • By Location:
    • Round 7 sucked in round1 but was fine in round2. Capitar is at fault for that one.
    • All locations that had pop locks in round2 were fine in round1.
    • I see absolutely no consistenty of a location causing pop locks.

Brad - if you need the saves of the results please let me know, I saved each game's turn 36. Thanks for looking into it. I highly recommend you peek at all Kingdom factions and Kraxis. Something is definitely amiss. :cylon:

[Edit] Oh, and Magnar too. Hell, look at 'em all! :frogboy:

Reply #13 from Elemental Forums Top

Oh man, I know I'm beating a dead horse here but ran a new game and Capitar died again...turn 13. Random large map, testing a mod. All the tests above were done with mods turned off so mods are not causing this. He refuses to found his capital oftentimes leading to untimely deaths.

Zip (last 3 saves) - http://dl.dropbox.com/u/19969679/Elemental/119n/119n_UDFv01i_13to14.zip

Reply #14 from Elemental Forums Top

Elemental is not that bad here, In Shogun 2, the Oda clan ALWAYS dies... on turn 2 :-(

Reply #15 from Elemental Forums Top

Updated 1.20 analysis...

Here is the map and the locations noted:

Here are the baseline saves (turn 1 of each round):

Here are the results:

Summary:

  • By Faction:
    • Capitar & Gilden plain suck. Pariden not so much but still unacceptable.
  • By Location:
    • Location does not appear to cause the problems.

Brad/Derek - not gonna beat around the bush. Extremely disappointed to see these fundamental AI problems made it into 1.2.

It's a good thing I'm white & nerdy.

Reply #16 from Elemental Forums Top

1.20a issue.

Playtesting 1.20a and UDFv02.11 - ran into my good buddies Capitar. Turn 33, pop locked.

http://dl.dropbox.com/u/19969679/Elemental/120a/120a_udfv0211_33.EleSav

Reply #17 from Elemental Forums Top

Quoting AlLanMandragoran,
1.20a issue.

Playtesting 1.20a and UDFv02.11 - ran into my good buddies Capitar. Turn 33, pop locked.

End of AlLanMandragoran's quote

 

Ooops. My bad. At least for Capitar, it's UDF's fault. They didn't have huts available :P. It's now fixed in the newest version.