Release Notes

Albert v6.0 Date (Mar 2013)

  • Requires AISever version 0.38 or later
  • Requires AIMapper version 0.41 or later
  • Capable of press levels 10, 20 and 30
  • Major Release with addition of Press for Orders during movement phases
  • Capable of press levels 10, 20 and 30
  • Capable of PCE, ALY, DMZ, XDO, NAR, NOT, DRW, SLO, AND, ORR

Albert v5.9

    • Requires AISever version 0.36 or later to operate
    • Capable of press levels 10, 20 and 30
    • Capable of PCE, ALY, DMZ, DRW, SLO, AND, ORR
    • Fixed bugs with Deceit module, and multipower ALY agreements
    • Modified code structure for human mimicing delays
    • Adjusted usage of DMZ and ALY.

Albert v5.8

    • Requires AISever version 0.36 or later to operate
    • Added press levels 10 and 20
    • Capable of PCE, ALY, DMZ, DRW, SLO
    • Capable of understanding multi-power and multi-province agreements.
    • Adjusted message flow control
    • Added delays to mimic human which can be disabled with -t parameter.

Albert v5.7

    • Added move Triangles to promote a Fleet or Army to a more valuable coast
    • Added penalty for leaving SC's exposed to allies.
    • Forms much stronger alliances from opening. Especially the high tension ones like ITA-AUS, TUR-RUS and ENG-FRA
    • Added bias for strong aliance and helps keep weaker ally from getting too weak. Reduces SC greed a bit between allies.
    • Further endgame lockup adjustments.
    • Fixed some bugs regarding grabbing empty SCs
    • Under certain conditions GER allows RUS into SWE.
    • Fixed internal handling of messages for press games with and without time limits.
    • Modified -d0 paramater, to run with variants like world3, runs fast like KMB with -d0
    • Neglects alliances during retreat phase for a power with only 1 SC
    • Modified convoy setup coding for provinces with low value. Helps ENG get army off the island.
    • Adjusted code for builds to better include value for open spaces next to target SCs

Albert v5.5

  • Overpowered SCs (tempi), is included in power size totals. Earlier identification of possible solo and neighbour threats.
    • Revamped openings distribution, and provided more consistency in moves and builds for the first few years.
    • Adjusted pressuring of allies code in opening moves.
    • Adjusted the risk of losing provinces at end game.
    • Gave ability to support 2 foreign powers in the same turn.
    • Adjusted Fear projection, reduced fear from home centers to allow easier alliance with close neighbours.
    • Reduced cohesion for late game, which better allows large powers to expand.
    • Provided better direction for units surrounded by very low value provinces.
    • Added code to determine if foreign power is applying pressure which may be precursor to stab.
    • End game SPR turn valuation is more conservative for alliance powers against power attempting to solo.
    • Alliance powers against solo will no longer choose low probability ordersets during endgame.
    • Changed Random number algorithm.
    • Fixed bug that gave too much value to defensive home center, for threat that was a couple tempi away.
    • Fixed bug with end game scoring for weak powers ordersets.
    • Fixed bug that hindered alliance formation in SPR turns.
    • Added code to help recently stabbed powers unite against common enemy.

Albert v5.2

    • Fixed some bugs
    • Improved recognition of future threatened SCs.
    • Improved ability to hold a position that could result in a successful foreign support near end game.

Albert v5.0

  • Added level 10 press. Uses PCE token
  • Completely reworked game openings. Albert now enters each game with a different predisposition resulting in an array of opening moves.
  • Gave Albert an algorithm to specifically analyze opening moves of all powers.
  • Changed Albert to hang onto alliances a little longer in certain circumstances.
  • Gave Albert a disincentive to pressuring allies in early and middle game.
  • Adjusted Albert to gamble positions less, especially during the endgame. He is more likely to hold existing positions or at least can better differentiate which moves could lead to losses. A bit more plodding in character.
  • Improved Albert’s willingness to offer a draw. He Will offer draw if map is static for over 3 years.
  • Fixed the 'obscure' adjudication bug.
  • Various other small bug fixes.

Albert v4.0

  • Modified code for alliance activities near end game. Initiates earlier organized attack.
  • Fixed bug, where Albert failed to recognize stab coming from a convoy.
  • Fixed bug which limited Albert considering foreign supports. It might overdo it a bit now but this also helps the endgame.
  • Introduced a dominated move algoritm for individual units to prevent moves like STP-FIN in S01
  • Modified code for late game as to which moves it considers threatening by Ally. More freedom given to allies to enter into adjacent spaces to SCs
  • More emphasis on protecting SCs late game over position. Will move to ally's unprotected SC to stop a solo. Is also tolerant of this when happening to him.
  • Improved efficiency of algorithm by 15%, but likely added 10% more things to do.
  • Many small fixes and adjustments.

Albert v3.0

  • Added Alliances for level 0 games (Albert will try to form alliances and build trust between powers. This is a major change over v2.1beta which essential is at war with everyone. Anyone who feels the alliance algorithms make Albert too easy to manipulate can use the [-n] command line parameter to force Albert not to form any alliances.)
  • Includes timer-delays for Adjustment/Retreat phases in an attempt to disguise itself as a human in RT games.
  • Added auxilliary provinces into the strategic evaluation. (Albert treats any move into a province that only borders his SCs as an act of war.)
  • Revised build Algorithms to properly handle multiple builds.
  • stalemated.

Albert v2.0

  • will only play a level 0 game.
  • Newly added -d[] command line parameter for 'depth of thought' 0-100 default runs at 50 ( at -d0 Albert will still still win over KMB by 10-15% )
  • Runs twice as fast while analyzing much more than version 1.0.
  • Revamped iteration control function
  • Iterates through each power 15 times trying to optimize its value.

Albert v1.0

  • Will only play a level 0 game.
  • Based partially on KMB version 6.0 code
  • Analyzes Position for every Power on the board.
  • Iterates through each power 10 times trying to optimize its value.
  • Calculates the probabilities of provinces being occupied for each iteration
  • Completely new orderset evaluation system based on probabilities of success.
    • Added a commandline parameter '-t' to put Albert in tournament mode to eliminate timer-delays.
    • Added penalties for moves A-B&B-C when A could move to C.
    • Albert will now offer draws, only when it is significantly trailing. If Albert is leading it will not offer a draw, even if
    • It is now possible to force Albert to offer a draw, by using the checkbox in the dialog window.
    • Added ability to support foreign units, including foreign self-bounces.