Does this tell me to dodge?
No. It gives you a probability and a band — really low, low, neutral, high, really high — for the lobby in front of you. What to do with that, including whether leaving champion select is worth it for you, is your call. See the dodging guide for more on when a low reading is actually worth acting on.
What does the percentage mean?
It's your team's estimated win probability, built from aggregate champion win rates for every recognised pick on both teams. It is not a measurement of the players in the lobby — see how it works for the full mechanics.
What counts as a "really low" lobby?
A reading more than one and a half standard deviations below 50% for your tier — which works out to roughly the bottom 1-in-15 lobbies at that rank, not just "below average." Full details, including the actual standard deviation for each tier, are on the how it works page.
Does it work before all ten champions are locked?
Yes. Every unpicked or unrecognised slot is simply left out of the calculation rather than guessed at. Because a partial reading is less certain than a full one, the app also widens the range it takes to call a lobby "really low" or "really high" until all ten picks are in — so a partial reading won't sound more extreme than the evidence supports.
Why did it misread my screenshot?
The app reads champion names with OCR, and OCR isn't flawless — low contrast, unusual crops, and especially compression artifacts can all cause a misread. If a pick comes back wrong, you can correct it by hand, which re-runs the calculation so the win probability and band both update to reflect it. If misreads keep happening on a screenshot you think should be readable, send it over and tell me which champions it got wrong.
Why is PNG better than JPEG?
JPEG compression introduces artifacts that blur fine detail, including the small text OCR relies on to read champion names. A screenshot saved as PNG resolves measurably more champions correctly than the same screenshot re-encoded as JPEG. If your screenshot tool gives you a choice, prefer PNG.
Do I have to enter a Riot ID?
No. Without one, the app scores your lobby using win-rate data and calibration bands pooled across all tiers, which is still a reasonable default. Entering one is purely optional and only ever used to narrow that default down to your actual rank.
What does entering my Riot ID do?
It looks up your rank so the app can pull champion win rates and calibration bands specific to your tier instead of the all-tiers default — which matters, because both the typical win rates and how spread out results are can differ noticeably from one rank to another. See the privacy page for how long that lookup is kept.
Is my screenshot stored?
No. The server checks it against basic limits — under 8 MB, PNG or JPEG, no more than 16 megapixels — decodes it in memory, reads it with OCR, and discards it. It's never written to disk and never sent to a third party. Details are on the privacy page.
Is this against Riot's Terms of Service?
The app reads a screenshot you took yourself; it doesn't interact with the League client, read game memory, or automate anything in-game. That said, this isn't legal advice, and Riot's terms can change — if this matters to your situation, check Riot's current terms yourself.
Why doesn't it know about counters or synergies?
Because it isn't built from that kind of data. The model works entirely from aggregate, per-champion win rates — it has no notion of how two specific champions interact on the map, good or bad, and no notion of lane matchups either. That's a deliberate limit, not an oversight: aggregate win rates are the one input the app can compute reliably, and it would rather be honest about that than fake a matchup opinion it can't actually back up. See about and how it works for more on that choice.
How current is the data?
Champion win rates come from recent ranked match data — which originates with Riot Games, though the win-rate aggregation itself is done by a third party — and shift with balance patches, so the numbers move as the game does. On the app's side, a fetched win rate is cached for up to 12 hours before it's refreshed, which keeps the app responsive without hammering its data source on every single lookup. The app doesn't control how often that source itself recalculates its figures.
Is it free?
Yes. There's no account, no subscription, and no paywall. See about for how the site is funded.
Why might a partial lobby look less dramatic than I expected?
Because fewer locked-in picks is weaker evidence, not stronger, so the app deliberately widens its bands as picks are still coming in. A reading that would count as "really low" in a full ten-champion lobby may only register as "low" with six picks locked — the model is being appropriately cautious with a smaller sample, not underselling the lobby.
What ranks does it support?
Iron through Challenger individually, plus an all-tiers pooled scope used whenever no Riot ID is entered or a rank can't be resolved. Each of those eleven scopes has its own calibration, because both typical win rates and how spread out results are can differ noticeably from one rank to another — see how it works for why Challenger's readings are naturally more spread out than Diamond's.