ko.lol vs Blitz
Blitz is a desktop app that imports runes and builds into your client automatically. ko.lol is a desktop app and a stats site that shows the measured games behind every recommendation: the sample size, the patch, and a 95% range. If you want setup with zero thinking, Blitz does that today, and it does it well. If you want to know why a build is recommended, how many games stand behind it, and when two builds are tied, read on. This page is the honest version, with sources at the bottom.
The short version
Blitz is the faster setup, and it does that job well. ko.lol is the one that shows its work: the games, the 95% range, and the ties. The rest of this page is the evidence for that sentence.
What does Blitz do well?
Blitz removes setup work. It reads your draft, imports runes and a build in one click or with no click at all, scouts the lobby, and reviews the game afterward. The convenience is real. For a player who would otherwise pick runes at random, the imported page is a genuine upgrade, because the import is usually a reasonable meta build. Our audit of League stats tools says the same thing in its Blitz chapter: the value is floor-raising, and honest framing would say so. Blitz also has far more match volume behind its numbers than we have today, and years of polish behind its app.
How does Blitz choose its builds?
Blitz does not say. It publishes no method, no sample handling, and no validation, so our audit had to infer behavior from what the app shows: the imports track popular high-win-rate builds of the kind every stats site serves. The marketing says the picks come "from Challengers", with no study and no numbers behind the phrase.
The deeper problem is what auto-import does to the data itself. Say the app writes the same Ahri build into a large share of Ahri games every day.
The import feedback loop
Why the recommended build's win rate always sits near the average.
-
1. The app imports one build
Every day, the same Ahri build is written into a large share of Ahri games.
-
2. The build becomes the data
After a few weeks, that build is most of the Ahri data, so its win rate has to sit near the average Ahri win rate.
-
3. The data says keep it
Every other build is left with a small, slanted sample: the players who rejected the import on purpose. Their numbers are noise with a slant, so nothing else is proven.
And back to 1. The import looks safe forever, whatever the meta does.
This loop also slows the tool down on patch day. When a patch changes an item, the old build's numbers cannot move until players stop importing it. Players do not stop, because the app keeps writing it. A system that always plays the favorite and never tests the alternatives does not find the best build. It finds a build and keeps it. That is why our audit calls Blitz the feedback loop with a download button.
ko.lol vs Blitz, side by side
| question | Blitz | ko.lol |
|---|---|---|
| What does a 52% win rate mean here? | A raw share of past games for that build, with no game count and no range shown. The build is most of the data by import volume, so the number sits near the average by the loop above. | 52 wins per 100 measured games, printed with the exact number of games, the patch, and a 95% range. A wide range means a weak claim, and the page says so. |
| Does it say when two builds are tied? | Not published. The import writes one page. | Yes. Comparisons have three verdicts: beats, too close, not enough games. Tied builds stay on screen as equals, and small samples are pulled back toward the champion average before anything is crowned. |
| Does it know who is playing? | Lobby scouting, yes: it shows ranks and recent games. Build advice: no published sign that it changes with your history on the champion. | Draft picks use your games on the champion and your recent form when the client reports the players. When it cannot see them, the note says it assumes average pilots. Never an invented player stat. |
| Where do the builds come from? | Not published. The marketing says "from Challengers", with no method shown. | Whole pages measured from our own match store: keystone, trees, rows, and shards from one page's real games, never stitched together. The formulas are on the methodology pages. |
| What happens on patch day? | The loop above: the old build keeps being imported, so its stats take days to move, and the app's users find out last. | The site rebuilds on patch day. The patch report shows what measurably changed, computed from games right before and after the patch boundary, with ranges. Thin reports are flagged unresolved instead of hidden. |
| Does it grade your actual game? | Yes: a post-game review with strengths and mistakes. | Yes: Build, Runes, and Play scores per player per match, each graded against the recorded pool for that champion and role, with the pool named. A thin pool shows no grade. |
What do you gain by switching to ko.lol?
- Numbers you can check. Every estimate shows the games behind it, the patch, and a 95% range. Thin data says "not enough games yet" with the games so far, instead of printing a lucky number.
- Ties called ties. Three verdicts only: beats, too close, not enough games. When two builds are tied, you keep both and pick comfort.
- Your games, graded. After each game you get Build, Runes, and Play scores, each one graded against the recorded pool for your champion and role, with the pool named on the screen.
- A buy answer computed at the moment you buy, from your champion, your items, your gold, and the enemy team. End-of-game item counts reward winners; buying is a mid-game decision, and that is where we measure it.
- Patch reports that measure what actually changed, from games right before and after the patch boundary, with ranges. No competitor publishes this kind of report.
- Free stats. No account, and no paywall on any stat page.
Where is ko.lol behind today?
- Match volume. Our store held 108,616 recorded ranked games when this page was built, and it grows by about 18,000 to 28,000 games a day. Blitz sits on far more games. Our answer is to print the count next to every number, so you can see the size of the evidence yourself.
- The pool. Our crawl starts from Diamond ladder players on EUW, Korea and NA and follows the people they play with, so the pool leans high. Low ranks and off-meta picks are thin, and the pages say so where it happens.
- Brand and content. Players have searched "blitz" for years. Almost nobody searches "ko.lol" yet. Our site is new, and it shows measured numbers with short written blocks, not a library of thousands of guides.
- Platform. The ko.lol app runs on Windows only.
How do you switch from Blitz?
About a minute of setup, then one game.
-
Download the installer
From the ko.lol download page. It ships for Windows today, and the page lists the current version and size (0.21.0, 83.2 MB, read from the release server when this page was built).
-
Open it
A short first-run setup leads into the full app: home with your profile and matches, the draft view, the in-game view, review, the tier list, tools, and settings.
-
Play a game
The app reads the game through Riot's own local APIs. When the game ends, your graded report is waiting.
No account, no API key. You do not have to uninstall Blitz to try ko.lol; run one overlay at a time at first, because we have not tested two overlays over one game. If you uninstall later, nothing is left behind: ko.lol never writes to game files, only to Riot's documented local interfaces.
Is ko.lol a good Blitz alternative?
If you want imports with zero thinking, Blitz already works, and switching costs you nothing but a download. The reason to try ko.lol is evidence. If you have ever wondered whether the imported build is actually the best one for your games, ko.lol shows the games, the range, and the ties. The website needs no install, so you can check a champion page before you download anything.
Why is Blitz's recommended build win rate always near the average?
Because the import volume makes it the average. When one build is most of the data, its win rate is pulled toward the champion's average by plain arithmetic, and every alternative is left with too few games to prove anything. Our audit documents this loop and its patch-lag cost in its Blitz chapter.
Does ko.lol auto-import runes like Blitz?
Yes, when you turn it on. Rune Auto-Set writes the measured page into your client through Riot's own rune interface. Tools that write to your client are off by default; you switch them on deliberately. The page it writes is the one with the measured games behind it.
Is ko.lol allowed under Riot's rules?
ko.lol reads only Riot's documented local APIs: the client API during champ select, and the live game API on your own machine during a game. It never reads game memory and never touches the other team's information.
Does ko.lol have as much data as Blitz?
No. Blitz has more match volume today, and that is the honest trade. Our store held 108,616 recorded ranked games when this page was built, and it grows by about 18,000 to 28,000 a day. We think a smaller store with honest counts beats a bigger one with none, and you can test that claim yourself: read the number next to any statistic on either product.
Can I use Blitz and ko.lol at the same time?
The website, yes. For the apps, run one overlay at a time at first. We have not tested two overlays over one game, so we will not promise they behave. You do not have to uninstall Blitz to try ko.lol.
See the math
Every number rule on this page comes from the same place: our methodology pages publish the formulas, the pool definitions, and the honesty rules, in plain words. Read how ko.lol measures everything. Watch the rules working on real champions: Ahri build, patch 16.14, Lee Sin build, patch 16.14, Jinx build, patch 16.14.
Use it in game
The ko.lol app shows the measured buy answer while you play and grades every game afterward. This page compared the numbers; the app is where they answer the question you have while the game is running.
Free. For Windows.