Factorio Forums

So, after banging at Wave Defense, I finally completed it. Around day 10 or so. Probably could easily do it by day 8 or even earlier I think.

It took a surprisingly (though very fun!) amount of time to figure out a winning approach.

I started with various failures trying to throw up defenses, like bullet turrets feedings bullet turrets; flame turrets; flame turrets and lasers; flame, lasers, and bullets. A well thought out base BP, at least for me, was a must.

The last approach with the entire kitchen sink backing me, I liked quite bit and was mostly effective - but still, I could only get to day 23+ or so and not really much to show for it. I was always so busy keeping the defense together I never got beyond the defending and eventually was overwhelmed by evolved biters. Plus all that laser power bursting!

So, as with all of my factorio play, after I had squeezed out all the thoroughly and very entertaining 'winging it gameplay', I took a step back. Nothing I was doing was really working and so I experimented and rethought my approach.

There were two ahah moments / tricks that were key for me, though I'm pretty sure now that I can win without the second one.

The first trick came from my frustration with the very very annoying personal robotport. I always kept it off except to mow down trees. I hand built my blueprint. Why? The little personal construction robots would always build off of them in absurdly inefficient order, drain their batteries at inopportune times and painfully, they flew too slow to follow me and ended up dragging way behind as I zipped across the base.

I actually had to stop several times for the robots to catch up, recharge and then re-enter the personal roboport to sleep for a bit! Stop? Sleep? In wave defense?

Anyways, hand building always seem to worked better. But still, I am very slow on the keyboard. So, I thought about it and finally did the very obvious thing everyone likely ends up doing - I split up my blueprints into a personal roboport optimized order. Wow. I was rather stunned how quickly the base came together. I had an artillery shell line producing by day 6, and that was just on the first run through without really much effort on optimizing my over all build order. I had been lazy and only split the base BP into 3 sections.

I've seen a few Any% speedrun videos and I imagine if I rewatched them, I'd just cringe as their bots randomly wandered off building things. My goodness folks, break up the BPs. Lay down power and belts first. The efficiency improvement will shock you.

The second trick*, may seem even more obvious, but honestly I never bothered looking at it as I was having fun trying to make my approach work without it. And really, it feels like a bit of a cheat. Yes, I'm talking about the upgrades. Once I had a few shells, I blasted a base or two and had the prebuilt turrets nicely upgraded. No need for a custom defense at all, just use the wall turrets that comes out of box as part of the scenario.

Maybe make upgrades an option?

Though TBH, I don't think it really matters and you can win without them. I was producing around 20 shells per minute by day 8 and I think you really only need about 160-200 shells or so to win the game. I am pretty confident I can accelerate the peak artillery shell production to day 6 if I wanted to by just splitting up the base BP into smaller, more optimized subset BPs. My bots were still doing a lot of dumb things with my 3 BP / day 10 win.

I've also found that you generally don't lose the game with zero new defenses until around day 9 when they overwhelm your base. Again, I'm pretty confident that I would have the map cleansed of the natives by then and just taking out the nearby bases stops the biters from coming.

Anyways, sorry for the long post, but the scenario was a lot of fun and wanted to share my personal victory. Fast game with rapid closure and a very clear build optimization problem.

Really hope WUBE releases more scenarios.

I think I'll try the rocket rush next, but this time I'm going straight to recursive blueprints with an optimized build order. For me at least, especially after how wave defense played out, I believe the recursive blueprints are the game at its best, and all the manual keyboard mashing seems a bit beside the point. Though maybe that's just my aching RSI wrists talking.

* Easiest win, as I learned from a reddit post after celebrating my victory, is just leaving the base early after setting up some quick smelting lines. Take out a few biter bases with turret creep to grab some low hanging upgrade points. Yeah, that looks like it'll work, though maybe not the fastest approach.

Interesting, when I build something using a blueprint I prefer it to be done with the personal roboport robots. Reason being that when I do so, I am usually some distance away from my base and really do not want to wait for the bots to come from the base and bring the various materials.

Yeah, the personal roboport can be alright if you're not under massive time pressure. If you grab the early upgrade points you can probably give yourself some breathing room of 4-5 days (about 30 minutes), or even more.

Here is an example of what I am talking about. Two production curves from ore/plate builds, one where a single BP is used, and one where belts/power/inserters are laid out first, and then miners/furnaces.

Both builds completed in the same amount of time (exact same material deployed), but look at the production curves taken at the exact end of the builds. The 1 BP approach had 206 iron plates built at end of build, while the split BP approach had 822. This type of activity can really add up, especially if the iron was used for another set of producers (say miners), as the bonus production from those already built producers will be added on top of any BP layout efficiencies used for those as well. That iron might be used for producers, such as assemblers and inserters. This compound benefit, when multiplied together, becomes quite significant.


https://imgur.com/a/dFZCc5N



https://imgur.com/a/opKQaz8

JIT manufacturing, baby!

I've found pretty much every optimized factorio build has found its way into a JIT steady state eventually, but few of the builds themselves are manufactured in this manner. The Speedrun guys, especially phoenix, are poking at this, but I don't know how much they are spreadsheeting it and getting the precise numbers. I do see his robots flying around haplessly though with poor priority order so I know there are definite inefficiencies.

Also, the keyboarding in speedruns likely plays a huge factor in their build orders, something I have no understanding or talent for.

blazespinnaker wrote: ↑Fri Oct 02, 2020 4:04 pm Yeah, the personal roboport can be alright if you're not under massive time pressure. If you grab the early upgrade points you can probably give yourself some breathing room of 4-5 days (about 30 minutes), or even more.

Or if I am building something using partial blueprints (I do not think Factorio has a way of editing a blueprint without building it first). Especially when I am further away from the base, since I get bored waiting for the robots to fly all the way there (that's assuming I have roboports built that far out).

I wish that, along with the "disable personal roboport" button there was a button "use only personal roboport", especially when I am clearing out stuff to re-do it (meaning - no blueprint, I am just trying things out to see if/how they work). Robots fly in, grab part of the objects and carry them to a storage chest at the other end of the base.

Yes, you can edit blueprints in that you can right click on objects to remove them from either the blueprint itself or from the list of objects that will be used (on the left hand side). You can't add new things to blueprint though, that's the only thing.

I agree, the confusion of roboport versus personal roboport causes a lot of problems, especially with the spidertron.

Have you tried wave defense? It's a lot of fun. That and rocket rush are really great. I'm surprised there isn't more discussion about them.