RepMastered™

Downloading replays and maps requires you to login. There is also a daily download limit based on your account type, summarized in the following table:

Name	Price USD1	Download limit2
★Free	$0	203
✧Silver	$5	200
✩Gold	$14	2,000
✪Platinum	$29	20,000
💎Diamond	$99	100,000
✨Infinity	$199	10,000,000

1 - Price is a one-time fee (there are no recurring monthly fees) 2 - The download limit is daily, resets after 24h of the first download 3 - Free download for unverified email domains is blocked

The ⚙️Settings page displays your current account type, used / available daily download quota and also when the daily quota resets.

Yes, there is. Check out the 🌀API doc page.

Currently "private" replays are not supported. All uploaded replays are public.
However, if you're willing to pay, I can setup, host and manage a private RepMastered™ server for you or for your clan. Contact me via email if you're interested.

You can do so by re-uploading them after logging in. All replays are added to "⬆My Uploads" automatically if you are logged in while uploading.

You may only upload replays that do not exceed 1 MB. Bigger replays will be rejected. Replays that have an invalid game date will also be rejected.

Statistics are calculated from the uploaded replays, so the accuracy is limited by this.
For example replays do not store the information who won. I wrote some algorithms to detect winners based on what's available in replays (see the How is the game result (winners / losers) determined? question below), but it never will be 100% accurate. I encourage to upload replays even if a game already exists, because having multiple replays increases the chance to more accurately detect winners (the system can detect if multiple replays belong to the same game). So the system and accuracy evolves naturally as more replays are uploaded.
Also, whenever I make improvements to the algorithms, I apply them retrospectively to past replays, so the win ratio, statistics and all other displayed information improves and gets more accurate over time.

Short games do not distort the statistics as games shorter than 110 seconds are already excluded from statistics (see Excluded Games column).

As noted above, replays do not store the information who won.
Winners are calculated using the "largest remaining team wins" principle. RepMastered™ determines the in-game teams, and tracks players leaving. The team that leaves last, or the largest remaining team is appointed the winner.
Notes:

• If there is a team full of computer players, the result will be marked "Unknown".
• Having a chat message in the game improves winner detection accuracy. So do leave a chat message in your games 😉

In general games with computers are included in matchup stats.
BUT if there's at least one team full of computers in a game, the game will be marked as "Unknown", and unknown games do not distort the win ratio (win ratio = wins / (wins + losses) * 100).
That being said, if there's a computer in a 1v1 game, that means one team is full of computers for sure, so it will not be counted in 1v1 matchup stats.
Similarly, if for example in a 2v2 game one team consists of 2 computer players, that will also be marked as "Unknown" and not distort the 2v2 matchup stats.

APM stands for Actions per Minute, it is a measurement of how fast players are during games, how many commands they issue during 1 minute. Higher APM means more activity, more frequently issued commands.
APM calculation is pretty simple: one just have to divide the number of commands by the number of minutes during which those commands were issued. For example 200 commands in 2 minutes results in 100 APM.
Often the first minute or so is excluded from APM calculation, but Blizzard does not do so, and so RepMastered™ doesn't exclude the first minute either.
However, the time used in calculation is the timestamp of the last action of players, so if a player leaves early, his / her APM will be calculated using the time while he / she was in the game.
Some external readings on APM: APM on Wikipedia, Actions per Minute on Liquipedia, Sc2gears APM Types

EAPM stands for Effective APM. EAPM excludes commands from the APM calculation that are considered to be ineffective such as:

• Unit queue overflow: training units that cannot fit into the queue
• "Immediately" (dt < ~0.8s) cancelling a unit or upgrade
• Repeating certain commands too fast (dt < ~0.4s) such as Stop, Hold, Attack, Set Rally etc.
• Changing selection or reselecting the same units too fast (dt < ~0.3s); note: double-tapping a hotkey group to center the group is not considered ineffective
• Repeating certain commands (without time restriction) such as Morph, Upgrade, Merge Archon, Lift off, Cancel Build etc.
• Repeating the same Hotkey Add or Assign command

Redundancy is simply the ratio of ineffective commands compared to the player's all commands.
Redundancy = ineffective commands / all commands * 100% = (APM - EAPM) / APM * 100%.
20% redundancy for example means that on average 20 commands out of 100 are considered to be ineffective.

The Map tab of the game details page can show "Hot points" and "Hot areas" in the background of the map view (as layers).
Hot points are basically the visualized target points of players' commands on the map.
The map can be divided into regions called Hot areas, and hot points falling into these regions are summarized. These regions then are visualized with a brightness proportional to the number of their hot points, relative to all other hot areas. The region with the most hot points is the brightest, a region with no hot points is completely transparent.
Hot points and Hot areas show you where big things like battles, intensive micro operations take place on the map.
The "Hot" window is a moving time window that defines which hot points to consider for the Hot points and Hot areas features. For example Last 1 min means to only visualize / include hot points that happened in the last 1 minute preceeding the selected / current time, which visualizes the recent battles and actions. You can increase / decrease this time window, and you can even select From start to consider all hot points from game start (up until the selected / current time). Animate the map to see how hot areas change!
Note: only hot points from selected players are considered. If you're interested only in a specific player, deselect other players.

Replays only store commands issued by the players, but not whether those commands are carried out successfully. For example a player may select a Probe and issue "Build Pylon" multiple times before the Probe reaches the destination where the Pylon should be built. There is no 100% way to determine which commands are unsuccessful (in order to filter them out).

If you want your tournament listed, contact me with the details. Replays must be provided (not necessarily all). See the ℹ️About page for my contact.

Yes, a 1.0 request / second rate limit is enforced by RepMastered™ with a 60-second time window.
This is sufficient for most clients, you will likely never hit this limit unless you're using a web crawler to fetch data from the site very fast.
Note that not all requests are rate limited (e.g. static resources and map images are not).
Each response to a rate-limited request contains an X-Ttl HTTP header telling how many seconds left from the current time window, and an X-Rl HTTP header telling how many requests you have left within the window.
If you go over this limit, your requests will be denied with a HTTP 429 response until the current time window is reset, and a Retry-After HTTP header will also be set to the time you must wait before a new request will be served.

Yes, you can, but be sure to set at least 1 second delay between requests to not overload the server.
Please see the Is there a limit how fast I can make requests to the site? question above for details.
In the future I might add an API so you can fetch data easier and more efficiently.
Also check out the 🌀API doc page.

The site is free for all with certain limitations. You can upgrade your account by donating to increase those limits. Check out the ⛑️Donate page.
You can also support me via GitHub Sponsors.
If you plan to use Digitalocean, you can use my referral link to register. You earn $200 and you also help me by doing so: https://m.do.co/c/7bdc66c1fe73
You can also help me by spreading the word about RepMastered™. Tell your friends about it.
And last but definitely not least: upload as many replays as many you can get your hands on 😉