Yes, some of it counts, but an hour of playing a game in your target language is a different unit from an hour of watching video, and it’s more honest to log it as its own thing. You usually hear fewer words in a play hour. In exchange you pay closer attention to each one, and the game shows you whether you understood. Below is how we’d count it if we were keeping a log ourselves, and where the evidence runs out.
How input learners count hours
If you follow Dreaming Spanish, you probably know your hour count off by heart. The Dreaming Spanish roadmap splits learning into seven levels tied to hours of comprehensible input, running from 0 hours at level 1 through 50, 150, 300, 600 and 1,000 up to 1,500 hours at level 7, where you’re an effective user of the language overall. The roadmap is careful to call these general estimates. It says related languages go roughly twice as fast and very distant ones roughly twice as slow, and that the quality of the input changes the pace too.
The progress tracker lets you add input from outside the platform, such as a Netflix series, a podcast or a YouTube channel, and those hours go into the same total. The Dreaming FAQ describes input as understanding words and sentences in context when someone is actually trying to communicate. It rules out grammar study, exercises and flashcards, and it suggests keeping reading as a separate count, in words read rather than hours. So a precedent already exists for logging a different kind of input on a separate line.
Refold draws a different line. Its roadmap separates active immersion, where you give content your full attention, from passive listening in the background, and some Refold learners log the two separately for that reason.
Both systems have helped a lot of people reach real comprehension, and neither one says anything about games. That’s why the question keeps coming up.
Why a play hour is a different unit
Picture an hour of a good comprehensible input video. Someone is talking almost the whole time, with drawings and gestures to help you follow. Every minute is packed with language aimed at a learner.
Now picture an hour in a game like the dungeon on mossyrune. You’re locked in with an old prisoner, a nervous merchant, a ghost and a guard, and you need to get out. Characters talk to you in the language you’re learning, but a lot of the hour is taken up by other things. You walk from one cell to the next, you stop to think about what the merchant meant, and you answer. On beginner and intermediate you can answer in your own language, which is fine for play but obviously isn’t input. You also replay lines and slow them down. Add it up and you’ve heard fewer words in that hour than you would from a video.
What you get in return is attention. When the merchant names a price, you listen properly, because you need the number, and when the prisoner asks what you found, your answer only helps if you followed the others. It’s hard to half-listen when a door stays shut until you’ve understood. Refold’s line between active and passive listening gets at the same thing, since it treats a minute of full attention as worth more than a minute in the background.
You also get a check. A video can’t tell whether you understood it, and you often can’t either, because nothing depends on the difference. In a game the result shows up. If you walk to the wrong place or offer the guard the wrong thing, you’ve learned something about your comprehension that a video wouldn’t have told you.
So the game and the video are measuring different things, and neither wins outright. Video gives you far more words per hour, with no setup and at any time of day, and for most learners it will stay where the bulk of their hours come from.
Crosstalk is interactive input already
The Dreaming Spanish roadmap calls crosstalk, where a native speaker talks to you in the target language and you answer in yours, the most efficient activity you can do at the start, if you can find someone to do it with. A crosstalk session has the same shape as an hour of play. The other person talks, you listen closely, you steer with a few words of your own, and they can see whether you’re following.
The games on mossyrune work in a similar way. On beginner and intermediate you can reply in your own language, and the character understands you and answers in the one you’re learning. You could think of it as crosstalk with a scene around it and something at stake. If you count crosstalk, you probably log the whole session rather than timing the pauses. A game has more going on around the conversation than a video call does, which is why we’d count it more carefully.
How we’d log it
Here are three ways to count it, from strictest to loosest.
Count only the listening. Add the minutes when a character was speaking and you were following, and skip walking, thinking and your own replies. It fits best with a roadmap built on hours of listening. Every playthrough is saved to your notes automatically, including the conversation, so you can look back at how much was actually said.
Log it on its own line as “playable input”. Keep your main total for video and audio and track game time beside it, much as the Dreaming FAQ suggests counting reading separately, in words. That way you don’t inflate your roadmap level, and after a few months you can see whether the weeks you played more lined up with anything changing.
Count it like crosstalk. If you already count crosstalk sessions in full, you could count a play session in full too. We think this overstates it for most people, but it’s a defensible choice if the game kept you listening for most of the session.
Whichever you pick, don’t count the time you spent reading hint translations or grammar notes as input. Going by the Dreaming FAQ, that time is closer to study than to input.
What we don’t know
Nobody has measured this. We don’t know how an hour of play compares with an hour of video for acquisition, and we won’t pretend to have a conversion rate. There are good reasons to think attention matters. Richard Schmidt’s work on noticing (1990) and Laufer and Hulstijn’s involvement load hypothesis (2001), which proposes that having a real need for a word helps it stick, both point that way. But neither studied games like these, and both are argued over.
It may turn out that a play hour is worth less than a video hour for raw acquisition, and more as a way to find out where you actually are. It may vary a lot by level. At the very start, a character who holds up the bread while saying “bread” is doing what a good CI teacher does. Later on, the denser input of video probably wins on volume. If you keep a separate log, you’ll have better data on this than we do.
Playable input
We call this kind of material playable input. Watched input carries on at the same pace whether you followed it or not. Playable input waits for you, reacts to what you do, and only moves on once the meaning has landed. It’s one part of what we mean by comprehensible gaming, games built so you can follow them where understanding changes what happens next.
If you want to see how it fits your own log, the games on mossyrune run in any modern browser, including on a phone. A free account comes with 20 credits, enough for one full game, and after that there’s a monthly plan or a credit pack that doesn’t expire. For more on using games alongside an input routine, read comprehensible input you can react to, or see how the games work.