A listening game on an answering machine. Each voicemail is read out loud by the device's own voice, then two or three questions ask about what was said. Botto the robot answers every question too, and gets some wrong.
Three messages a day at three levels; at levels two and three, some questions ask what the caller means rather than what they said. The first run finished each day is the official one: it keeps the streak, the level best, the daily record against Botto and the Gold Tapes. More runs that day are practice and leave those alone, though every finished game still shows on the scoreboard.
Seven sets of ten voicemails with two questions each, and a Harder Round in every set: ten longer messages with three questions each. A Harder Round opens when its set's ten are done, or straight away with the Harder Rounds setting. Some callers are scams or sales calls.
Voice: Slower reads everything more slowly. The speed buttons fine-tune it from 90% to 110%, the volume has five steps, and the big button mutes everything. Voice check shows the voices on this device and lets you pick one for women callers, men callers and the daily messages.
Every answer is saved in this browser under the client code: the question, the answer picked, how many times the message was played, and how long the answer took. Records can be saved as a CSV file (one row per question) or as a backup file, and a backup can be restored on another device. Nothing is uploaded. Some browsers send spoken text to an online voice service when an online voice is chosen; Voice check says which voices work offline.
Every question has four answers, so guessing alone lands near one in four. The records say when a score would be unusually high or unusually low for guessing alone (a one-sided binomial test at the 5% level for each). Below three questions, knowing and guessing can't be told apart.
Guessing at random is not the only way to score without listening. Across the whole machine, tapping the longest answer every time would get about three in four right: fewer on the daily machine's level 1, far more on the scam and sales tapes. So the records also show a longest-answer baseline for the exact questions asked. A score near that baseline may come from reading the answers rather than from hearing the message.
Questions on scam and sales tapes test judgement as much as listening, so the records count them apart. Between-the-lines questions are marked on the daily machine only; the questions in the seven sets are not marked either way.
It is a practice log. There are no norms, no cut-offs and no severity ratings, and nothing here is a test or a diagnosis. Reading it is your call.