archive by month
Skip to content

why Steamhammer needs a tactical analyzer

This picture of Steamhammer versus XIMP by Tomas Vajda is a perfect illustration of why Steamhammer needs a smarter tactical analyzer.

Steamhammer retreats into the enemy defenses

The hydras decided that they had to retreat from the carriers at the top. Where to retreat to? “Hey, look, those scourge have found a great location! It’s a safe spot, it’s close to the enemy base—let’s go there!” It follows a heuristic of trying to keep its units together. It hardly needs saying that the entire hydra force was lost trying to run past the cannons. In big army games, Steamhammer makes similar blunders nearly every game.

The poor scourge play is another issue....

The picture also illustrates the silver lining to Steamhammer’s excessive retreating weakness in this version. Because Steamhammer now plays weakly when it is locally behind, it lost games playing the merely strong 3 hatch hydra opening that it had settled on in the previous version. It was forced to find an even stronger counter to XIMP, 3 hatch before pool into hydras. Look at the worker counts—even after losing this army, zerg is still ahead because its massive economy can replace the losses while protoss will remain unable to expand. Thanks more to XIMP’s objectively bad strategy than to Steamhammer’s play, and despite the severe blunder, Steamhammer put the game away easily.

Trackbacks

No Trackbacks

Comments

Antiga / Iruian on :

I've mentioned to Dan a few times that a modern reinterpretation of ximp using Purplewave as a base, has quite a bit of potential. It'd be interesting to see someday.

Jay Scott on :

I’ve had a similar thought. Many bots struggle against the carriers. But I’ve noticed that most bots which make carriers also struggle to use them well. XIMP is usually adequate, and Tscmoo knows the micro techniques, while other bots tend to be shaky in their carrier skills. It would be fun to polish up Steamhammer’s carriers and write a XIMP-like strategy, which should obviously be called Ximpalong Cassidy. Defending with fewer cannons and more units would add flexibility. I’m not doing it soon, though.

Arrak on :

I think I had my scourge ignore combat sim to some degree, and my ground units ignore the location of air units when considering retreat. Not the best solutions ever, but they were quick --- though, with ground-unit runbies and doom drops, this type of problem would probably resurface. Anyway, I plan on rejoining the bot development soon. I'm considering posting the code on my Github, if you approve, and I'll carry on the Steamhammer spirit with an MIT license.

Joseph S Huang on :

You don't need permission to use MIT license...

Jay Scott on :

Just so. Post what you like. I put the license on in part because I don’t want to be a gatekeeper.

Joseph S Huang on :

How about if unit is off cool down and retreating but there is an emeny in range fire at it first and then retreat.

Jay Scott on :

That makes sense most of the time for reavers, which run away slowly and do heavy damage. Mostly, units have to watch for timings when they can do some damage and still get away. It would be a valuable skill.

Dan on :

Not a universal answer. Consider the case of Dragoons retreating from Dragoons -- you can usually get away if you just run, but if you fire back you will eventually die.

PW has multiple styles of retreating and weighs them based on contextual factors.

Joseph S Huang on :

It might not make sense for a badly damaged unit, but otherwise i don't see a downside.

Jay Scott on :

Retreating vultures can shoot at chasing zealots because vultures are faster and have more range. Now consider vultures fleeing from dragoons: It might make sense to turn around and fire a shot if you have gained some advantage (you’ve reached the top of a ramp, or one dragoon is separated from the others)—but you are retreating because you will lose a stand-up fight. Usually, you will come out behind in a brief skirmish too. The dragoons will shoot before the vultures get back out of range, and do more damage. Dan’s example of dragoons chasing dragoons is clear too: You are retreating because the chasing dragoons are stronger, so if you stop to fire a shot, you will usually be hurt more.

Dan on :

Where it gets interesting are the marginal cases in the middle.

Dragoons can't outrun Vultures. So unlike the Dragoon vs. Dragoon case, "just run away" is a bit less appealing of an option. Dragoons also deal a lot of damage per shot and don't have to turn, so for them it can be favorable to take shots while retreating. Zealots vs. Zerglings is a similar case. For both, context matters -- is there a bastion of strength that you'd want to retreat to in a hurry? Or are you making the best of a hopeless situation by taking shots while retreating?

Vultures trade poorly in slugfests against Hydralisks, and can safely run away from them, but because Vultures deal more damage per *shot*, kiting can be profitable depending on the numbers. -- if the Vulture stays out of range except when trading, it doubles its DPS relative to the Hydralisk. Dragoons vs. Marines is a simlar situation -- the Dragoons deal way more damage per shot, so it can be better to kite backwards than to retreat all at once and stand ground.

Dan on :

Every time you stop and shoot, you get shot back, usually moreso because they have more Dragoons. It's an even trade at best but usually a losing trade.

Add Comment

E-Mail addresses will not be displayed and will only be used for E-Mail notifications.

To prevent automated Bots from commentspamming, please enter the string you see in the image below in the appropriate input box. Your comment will only be submitted if the strings match. Please ensure that your browser supports and accepts cookies, or your comment cannot be verified correctly.
CAPTCHA

Form options

Submitted comments will be subject to moderation before being displayed.