GamesBeat covers Basic Combat Training
GamesBeat covered Basic Combat Training — ReBlink’s first public loop for collecting player training data to build AI agents inside a real strategy ruleset.
Basic Combat Training was not only a game launch. It was ReBlink’s first production loop for collecting player training data to build AI.
In December 2022, GamesBeat covered that launch. The public story was a new strategy title and an early audience. The internal reason mattered more: put a real ruleset in front of real players, capture how they play, and use that signal to train agents instead of waiting years to invent behavior in isolation.
That is the through-line from Basic Combat Training to the systems we build now.
Most studios still separate “ship a game” from “train an AI.” We treated play as data from the start. Matches, decisions, and competitive pressure were useful twice: as entertainment for players, and as supervised and implicit labels for models that had to act inside the same legal action space.
GamesBeat noted the franchise framing and the push to get feedback into the cycle early. What we want remembered is the mechanism underneath. Early public play was how we started building AI that could eventually plan, co-play, and be evaluated against the real engine rather than against a toy proxy.
That bet became Arena Tactics, Playthrough, and Playable: language control, agentic simulation, and player-facing modes that reuse the same intelligence. Basic Combat Training was the first time we ran the loop in public.
Read the GamesBeat coverage: