From Liquipedia StarCraft Brood War Wiki
This article is a Bot stub. You can help Liquipedia by expanding it.
[e][h]TerranZerg Halo
Bot Information
Hao Pan
BWAPI Version:
No wrapper
Terrain Analysis:
Iron bot
Bunker Rush, BBS, 2 Fac pressure, 14 CC

Halo is a StarCraft AI, written by Hao Pan. It uses BWAPI as an API for interacting with Starcraft: Broodwar.

Based On[edit]

Halo is a fork of Iron Bot.

Starcraft/BWAPI-related 3rd-party Libraries Used[edit]

AI Techniques[edit]

  • Bayesian Bandit‡
  • Continual Online Evolution Planning (COEP)
  • Potential Field
  • etc...

‡ See also: the author's paper published at the AAAI-20 Workshop on Reinforcement Learning in Games


  • On SSCAIT, Halo is named "Hao Pan" playing as Terran and "Fresh Meat" playing as Zerg (currently disabled).
  • Some consulting by Nathan Roth.


Playing as Terran[edit]

Halo utilizes an array of strategies for each matchup, ranging from rush to macro ones. Notable ones include:

† See also: Jay Scott's StarCraft AI blog post, Hao Pan's bunker rush

Playing as Zerg[edit]

Halo plays as Zerg under the name Fresh Meat. The initial development was based on Iron bot. However, it turned out that its crash rate was too high on BASIL (around 20%) which discouraged further development and resulted the author disabling it. Starting from 2019/05/30, a new Fresh Meat is enabled on SSCAIT and it is based on author's Zerg bot for AITT S2. It's quite simple at this time, as it only utilizes a few techniques such as Bayesian Bandit and COEP. Its performance on BASIL will be closely monitored and will serve for benchmark purposes.


Note: all results of SSCAIT are standings/winrate after the round robin phase
In Tournaments
Date Place Event Race Result Category
2017/2018 C171st SSCAIT 2017/2018 Winrate: 12.34% Mixed division
2018/2019 A512th SSCAIT 2018/2019 Winrate: 71.83% Mixed division
Feb/2019 A33rd AIST S2 0:2 Velocirandom Combat History: WWLL
2019/2020 A44th SSCAIT 2019/2020 Winrate: 85.23% Mixed division
2020/2021 A55th SSCAIT 2020/2021 Winrate: 82.73% Mixed division