We at Spawning Tool have been thrilled by the engagement and enthusiasm from the community in using our site. So far, we have had tens of thousands of replays uploaded and over 200 build orders shared. However, the biggest limitation and most common request we have is that you can only access the data from the browser. To address that, we’re excited to announce the Spawning Tool Build Advisor, the best way to learn a build order.
Through Overwolf, the Build Advisor allows you to follow a build order live with an in-game overlay during your game. Browse through the builds listed on our site, then select your winning strategy to use during any ladder or custom game. As soon as the game starts, the overlay will sync with the in-game timer and walk through each step of the build so you don’t miss a thing. And that’s it.
To get started, download and setup Overwolf, then install the Spawning Tool Build Advisor app. It will automatically launch when you start StarCraft.
We hope you find the app as useful as we have. As always, if you have any feedback, please reach out to us on twitter @spawningtool or via email email@example.com . We would love to know how we can improve your StarCraft.
SpawningTool holds the largest collection of StarCraft 2 pro-level replays, build orders, and replay statistics. The site uses machine learning and the information stored in the replays to generate these statistics, label build orders, and organize games by series/event.
With the success the site has seen over the last 6 months, we have no one to thank except the community. In order to give back, we are going to host a $200+ invitational and release a new plugin coming soonTM. This plugin is going to allow you to have any build from the site shown live, in-game, without disrupting your perfect micro of course!
Matcherino is a extremely powerful website that gives any tournament or showmatch host the ability to crowdfund their tournament. They recently launched their new Tournaments Feature, which has automatic payouts, crowdfunding capabilities, stretch goals, and more! We at SpawningTool are very excited to have Matcherino as a sponsor plan on inviting them to any other events we host in the future. We are going to be using this Tournament Feature in this Invitational
This tournament is also sponsored by OSC, providing OSC points to all participants!
The point distribution is as follows:
1st: 50% of prize pool, 80 OSC points
2nd: 30% of prize pool, 50 OSC points
3rd: 15% of prize pool, 30 OSC points
4th: 5% of prize pool, 15 OSC points
5th-8th: 5 OSC points
I usually try to post interesting things from Spawning Tool data, but here’s a very boring result from some analysis I did. Using the ~15,000 professional HotS replays in Spawning Tool, I bucketed them by month and then calculated the average game length. I was thinking that maybe game lengths would have increased as players got better and didn’t lose as much early. As you can see that didn’t happen as game lengths remained roughly the same (about 16 minutes)
I didn’t dig too much further because I didn’t really see any trends. You can see the data I used to generate the graph at
You might be wondering: what type of all-in starts with economic cheese? The type that wants to have a sustained attack as soon as possible, that’s what.
In this game, Snute cuts corners and goes for almost pure Drones for the first 4 minutes of the game. This allows him to get up to 6 Extractors and spread Drones across 3 mineral lines before he lines up his attack. From there, he streams Roaches and Ravagers across the map to bust down walls and forcefields to force his way into the Protoss base.
On the flipside, State played it safe opening Gateway first, then getting his Nexus afterwards. WIth a quick +1 attack, he also gets his 3rd base quickly after scouting Snute’s greedy build, and then into Blink.
When the Ravagers attack, State already has his simcity in place with Gateways clogging up the entry into his 3rd base, so despite being down at times 70 to 15 army supply, Snute has difficulty finding space to push in. State continues to build pylons to power his Gateways and for the Overcharge.
With Blink up, though, State makes a great hold: his 3 base economy (with more Probes than Snute has Drones) allows him to slowly get back into the game while Snute’s slow Roaches walk across the map. The battle continues on for minutes, but State keeps his 3rd base alive and evens the supply count. When the pressure begins to lighten, State is able to push for a 4th base as his minerals begin to run out his in main. When the supplys even up, he traps Snute’s army and pushes across for the win.
MorDka shared his 2base immortal adept all-in as a new way for Protoss players to remind Zerg players what WoL felt like. You can read up on how it works, follow the build order, run the real-time build order, read the analysis and watch replays about how it works. Here are a few thoughts from the creator:
1. How are you liking the LotV beta so far?
Legacy of the Void is a lot of fun and frustrations since it’s so different than HotS. The game allows for more comebacks with the new units and the stronger defenders advantage. The fast pace of the game makes it more stressfull especially for lower level players and that’s why archon mode is so good, I really enjoy it.
2. How well is the build working on the ladder?
Well… Actually after the newest patch the build doesn’t work so well. Slowing down the warp in time made the warp prism way more vulnerable to corrosive bile and the adept nerf made the build quite a bit weaker but i think that it’s still viable on master level.
3. What builds are your opponents using that are giving you the most trouble?
The game has a bigger defenders advantage right now so there aren’t any specific build orders that i have trouble with. It’s usually a specific new unit that causes the troubles. For example Liberators or Lurkers.
Of course, LotV isn’t coming out for another month, but we actually just saw our LotV beat HotS in our google analytics data, so it seems that the tide is turning, and we want to make sure that the most useful data is available easily. We will, for the foreseeable future, continue to keep our HotS site online, but especially with the replay format changes, those replays just aren’t so useful anymore.
One question we have had a few times is why we split the sites. The primary reason is because the data is very different, and we had to have separate logic for both of them. Specifically, the unit types and game speed actually require different backends in place. Because of that, we felt that the data was altogether incompatible, so keeping them together would just muddle up statistics.
As for why you don’t share you user account between the two? That’s mostly laziness. The 2 sites are on different databases, and I didn’t put forth the effort to sync up the authentication between them.
These histograms show the research time for various upgrades in Legacy of the Void. The data comes from ~2200 replays uploaded to http://lotv.spawningtool.comof all levels of play from the beginning of beta. About 25% of games are from professional tournaments. The same includes about 1500 Zerg replays, 1200 Protoss replays, and 1000 Terran replays.
The graphs are bucketed by LotV (real-time) seconds. In some cases, extreme outliers to the right have been removed from the data for a better visualization. Note that the Warp Gate timings may not be exactly right while Blizzard tweaks with the exact impact of Chronoboost, but it should still be very close.
If you want to play around with the data yourself, you can see it at
Let us know if there’s anything else you want to see from our data. We will try to see what we can do, but in the mean time, play around with the research tool to see what you can find yourself! http://lotv.spawningtool.com/research/winrates/
Also, if you could help us make the data as accurate as possible by uploading your replays, that would be amazing!
Over the past few years, Spawning Tool has grown a ton. Thanks to tournaments organizers and casual players, we have accumulated over 30,000 StarCraft 2 replays with 13 million build steps and 400,000 tags. Although the site can generate interesting statistics and predictions from the data, it can be daunting for players to find strategies to actually use in-game. You have to dig through hundreds of replays and hope to find a good, clean build order. We wanted to make that easier, so we’re releasing Spawning Tool Build Orders.
What do you do if you see a cannon rush? Maybe you cancel that base and go to the 3rd base location instead. Maybe you pull a few workers to try to poke it down in time. Or if you’re hydra, maybe you just pull the boys and go for a worker rush.
hydra goes Pool First against a Forge Fast Expand. His first expansion is to the gold base, but his Drone crosses paths with a scouting Probe. Seeing no natural expansion, the Probe turns around to cannon rush the gold base. Then, hydra pulls all except 3 of his workers with 6 Zerglings in production to go for the all-in.
Note one clever moment when hydra pulls back his Drones to wait for the Zerglings. Perhaps even more importantly, he saw the Probe leave the base, and that Probe could have potentially scouted his Drone rush with enough time wall in. By hiding them in the corner for a few seconds, he keeps the surprise on his side.
One more thing: hydra has his Drone mineral walk into his opponents base. This is quite important as this would allow him to walk past a “hold position” Probe into the base to attack a Photon Cannon. Stardust wasn’t walled in at all.
I wouldn’t recommend this as a planned opening, but keep it in mind as a response to a cannon rush. The game is going to degenerate quickly anyways, and I guarantee that this all-in will be quick.
Group D was the group of death: the reigning SSL Champion (Classic), the reigning GSL Champion (Rain), MVP’s top Protoss player (YongHwa). And Snute came out on top with an aggressive heavy Hydralisk style in ZvP.
Snute opens with a safe 15 Pool, 16 Hatch to avoid a cannon rush and transitions into a relatively normal 3rd base timing around 4:00 and Zergling speed soon after. Around 7 minutes, he goes up to 30 Zerglings to deny Rain’s 3rd base and immediately goes into Hydralisk production. He uses those to destroy the 3rd base again and create a soft contain. The safe opening allows Snute to apply pressure before typically 2 base timings come in from Protoss.
From here, Zerg will typically transition into either Infestors to lock down the Stalkers or more Vipers for the Blinding Cloud, but Snute gets away with building only 2 Vipers. Most of us don’t have army control like Snute, but he pulls it off with smart positioning and baiting forcefields. Keeping the Viper and Infestor count low leaves more gas for Hydralisks.
There are a few caveats to this style. First, this build is likely a response to the prevalence of Blink Stalkers in the current meta. Against a more traditional Colossi army, the mid-game pressure timing is much tighter. This seems like the biggest potential risk to this style
Second, there are different ways to transition. In his other games, Snute also opted for Roach Zergling pressure early on, so there are options for staying aggressive.
Third, Snute delays upgrades for a long time. His first attack/armor upgrade is +1 missiles at 12:00. This might be a risk against a double Forge style with more armor upgrades to buffer against the Hydralisks. To sustain the attack, you probably will need to go up to 4 bases. Snute gets it very late in this game, but you can see other timings in his games against Classic
Fourth, the late-game composition is presumably more flexible for how the Protoss responds. Against High Templar, Snute opted for more Roaches. If he had seen more Colossi, I would guess you would get more Vipers for the Abducts.
So in summary, safe opening + big Speedling attack + mid-game Hydralisk contain = devastated GSL/SSL champions