Harvesting: or How I learned to love the collector

Homeworld Remastered Collection Homeworld Remastered Multiplayer
#1

The following test and observations are brought to you curtesy of guru labs.

I’ve spent some time running test with harvesting units from the Hiigaran (HW2) and Taiidan (HW1) races to both get a feel for what is optimal harvesting and to compare the income potential of the two races. Below are the results of my testing, but the TLDR; is that they are surprisingly balanced in terms of income potential up to 10 collectors, beyond that the Hiigaran start to break away. Oh and collectors are morons.

Some facts and rough estimated times.

HW1 Collector
load - 500 ru
Speed - 110
Cost - 600 ru
Build time - 60 seconds
Collection time on rock - 40 seconds
Drop off animation time - 15 seconds
Hyperspace capable - no, but was in HW1 (bug or intended?)

Resource Controller
Research required - Capital Ship drive (Research ship + Cap ship research = 1550ru)
Built from - MS only
speed - 100
Cost - 900 ru (minimum 2450 for first, 900 each additional)
Build time - 1:10 ( minimum 3:40 for first, 1:10 each additional)
Drop off pads - 1
Hyperspace capable - yes

HW2 collector
load - 200 ru
Speed - 276
Cost - 400 ru
Build time - 30 seconds
collection time on rock - ~25 seconds
drop of animation time - ~ 7 seconds
Hyperspace capable - no

Mobile refinery
Research required - none
Built from - Carrier, MS
Speed - 226
Cost - 800 ru
Build time - 45 seconds
Drop off pads - 2
Hyperspace capable - yes in range of module

Findings. All test were done on Kharam Wreck using the closest 2 patches.
Controls included not utilizing the carrier in all but two test, which will be explained later. I also did not repositioning the mothership from her starting position. I did move the carrier out of the way in the case of the Taiidan due to it being closer to the starting patch then the Hiigaran. The incomes below shown should not reflect the potential because in a real game, players would utilize the carrier as an extra/early drop point and reposition ms for faster delivery.

Test 1 - 6 H1 collectors vs 6 H2 collectors (pretty much just hit H as they came out and walked away)
Results:
H1 - Reached 10k ru at 7:48 seconds
H1 - Had 18.5k at 15 minutes

H2 - reached 10k ru at 8:24
H2 - Had 16k at 15 minutes

Winner, HW1 collectors

Test 2 - 6 H1 collectors + controller vs 6 H2 collectors + mobile refinery

Resulting total ru collected
H1 - 22500 - 2450(total 1st controller cost) = 20k
H2 - 21k - 800 = 20k

H2 has caught up at this point

For the next 2 test, I allowed the second patch H1 collectors to drop at the carrier that I moved out of the way due to the amount of time it takes to get a Controller out. Once the controllers were out they both were moved to the patch. For H2 collectors this was not needed due to mobile refineries.

Test 3 - 10 H1 collectors +2 controllers vs 10 H2 collectors + 2 mobile refineries (harvesting 2 patches)

Resulting total ru collected:
H1 - 33.8k - 5750 = 28k
H2 - 33.2k - 3200 = 30k

Winner HW2 collectors

Test 4 - 14 H1 collectors + 2 controllers vs 14 H2 collectors + 2 mobile refineries

Resulting total ru collected:
H1 - 40k - 8150 = 32k at 15 minutes
H2 - 40k - 4800 = 35k at 15 minutes

Winner, HW2 collectors

Observations:

  1. I must admit I was surprised at how well the HW1 collectors did. Before running these test I had timed the collection time and the drop off animation time. Combined with the fact that controllers only have one pad I pre concluded that the HW1 collectors would spend way more time then they did waiting on their turn to harvest or deliver. The results show that in a controlled environment this is not the case. Not to down play a 3k difference in the last test, that is a free destroyer to HW2 players, but it is closer than I originally assumed.

  2. Collectors are morons. They spend a fair amount of time waiting to harvest a rock that is already being collected even if there is an available rock to harvest, and I cannot figure out why. They do switch rocks from time to time (I first thought that they were sticking to the first one they touched) and it’s not a matter of switching to the one that is closest to them after drop off (I also watched one stand by idle waiting on a further rock to clear when a clear one was closer) But the AI could use some tweaking.

  3. The Hiigaran have a clear advantage in terms of ease of use and initial setup of their harvesting operations due to being able to build refineries from the start. This immediately frees up the carrier, mothership, or both for tactical movement. They are also better suited to relocate due to both collectors and refineries having double the speed. They can be built from both a carrier and ms and do not tie up the super capital ship hanger when deploying. And to add insult to injury the HW2 mobile refinery is packing.

  4. The HW1 collectors appear to be more marathon runners. I believe if I had allowed the test to run longer than 15 minutes, in both the 10 and 14 collector test the H1 collectors would’ve caught up. In subsequent test I plan on recording the times when overtakes happen.

  5. I am still gathering results to find the optimal number for each race per patch given the number of small/large rocks per patch. I.e. If I have 12 collectors between two patches, 1 patch a 2/1 small to large, 1 patch a 3/1, would it make more since to distribute my collectors 7 to the 3/1, 5 to the 2/1, or keep them 6 and 6. I’m betting this data already exist given this should be no different from HW2 and there were plenty of skilled folks that played that game. The limitation on how many collectors can harvest a rock at once is new to me (this limitation was not in HW or HW:C).

  6. issuing a guard order from the refinery to more than one collector do not keep the refinery/controller equal distant between the 2 collectors, the controller will only follow 1 unit.

2 Likes
Resource collector logic
HW1 collectors > HW2 collectors (MP whine)
#2

Nice work rata.
Thx.

#3

I’ve noticed that some, but not all, Taiidani collectors tend to go idle at some point. They sit on the asteroid with a full load, doing nothing.

#4

It is a bug, happens with kushan too, i recorded this:

I found out that it seems to be some type of colision bug, the more stuff you have nearby the more often it happens, try to keep the mothership and the carrier not too close of the asteroid field and you should be fine.

#5

You moved away just before I could see if my theory was right in the case of your recording.

What is captured here is similar to what I was talking about when a collector would sit an wait on a spot on a rock to open up, even if there was an available rock to harvest, and even if that available rock was closer than the one he grew fond of.

This isn’t just the HW1 races. I noticed this with Hiigaran also in both queuing up for harvest and queuing up for drop (what looks like happened here)

My theory is that somehow, Collector A is claiming it’s spot in the queue a particular pad or rock ahead of schedule, and when Collector B comes around to what would be a reasonable turn on the pad or the rock, since Collector A has already claimed that spot, it will just sit and wait for it to finish regardless of where Collector A is in the cycle.

Where this gives the appearance of a stuck collector. I’ve seen it get backed up to a point where a Collector B came of the pad and stopped, Collector A came to drop a load behind Collector B and then proceeded to Collect at the rock that Collector B was waiting on to free up, only it wasn’t taken when he originally came off the pad. Somehow, Collector A claimed it before dropping his load.

This IS happening with Hiigaran like I said, you can actually see it happen quite often, only it’s not as pronounced because of the short amount of collection time and drop of time the collectors spend. It’s very apparent when watching HW1 races due to the amount harvest and dropoff time, In the example I described above, that collector would’ve sat still waiting on that rock for a minimum of 55 seconds (harvest time + drop off time) plus the travel time between rock and refinery. In 1 minute you have a lot of opportunities to notice a collector taking a break.

Like I said in the OP. Collectors are morons.

1 Like
#6

Here’s some more stats from the files:
2000 HP on HW1. 6000 on HW2. HW2 also have like 10% armor at least on one side, I forget.
The mining rate is 8/s for HW2, 11.5/s for HW1.

This is why HW1 starts out so fast, then drops off. They both start with 6 miners, but HW1 ones collect 45% faster.
Some of that extra collection speed is lost in how slow they are.

Then they fall off because the 45% faster mining isn’t enough for all their drawbacks.
Their build time is much slower as well so your 14 vs 14 example is much worse given that HW1 races take twice as long to get 14, no?
Even if you accounted for that, more is lost getting a mobile refinery out.

I modded it and put it at 13 ru/s and dropped the starting count to 5, and that drops their initial mining by 8% but after you make 3 collectors you start gaining more.
The other thing HW1 races need is mobile refinery build-able at the start.

#7

Another option, similar to the argument folks are making regarding strike craft, is to give us all the same collector/refinery. Tweak the graphics to make them look similar to the old if you like. Get creative. But harvesting is something that should be identical between races.

#8

Eh not necessarily.

There is different research costs between races, different module costs, different types of units. They aren’t symmetrical.

I like the different collectors. They just aren’t balanced.

2 Likes
#9

oh I’m a fan of diversity. Just not sure resourcing is a place to try to mess around with balance. with collectors and controller/refineries at different loads and different cost, there isn’t really a way to ‘balance’. At one point or another one will always have an advantage in the test I showed above. I’ll admit you could call this a strategic element, just not sure if I think it’s a good idea.

Can you think of another example of an RTS where resourcing is different between army types?

#10

HW1 should be able to get a controller without researching for it, it is a huge tatical advantage for hw2

#11

Agreed, possibly. We didn’t have that advantage in HW1, we had to research, however on every map we started with a research ship, and on many defaults we started with a controller.

A BIG thing we are missing that I would like to have back is the ability to jump collectors. I like to believe that this was a conscious choice to take away, and not a limitation of the game engine. In HW1, collectors acted more like frigates instead of corvettes. IIRC, you couldn’t dock a collector, if you gave it a dock order, it would drop its load (I have to go back and check this).
Giving us jump ability back would give us back the ability to expand quickly or strategically. Plus allow us to protect us from collector sniping that the HW2 races have so many options to do at all points in the match.

1 Like
#12

Once they start making balance changes I would support this. Kind of a pro/con thing. Con you have to research the controller and it doesn’t have weapons like the HW2 ones and you cna only build it from the MS but on the pro side it can hyperspace and is way beefier and can repair fighters/corvettes and maybe some other things?

1 Like
#13

May as well make hyperspace and repair a research if it’s going to have that… Not make it delayed and put HW1 races as a disadvantage.


The not needing research to make it is one of the changes I put in my rebalance mod, here.

#14

Assuming the collectors weren’t such moron things aren’t nearly that bad.

I’d accept controllers requiring research but if we were given one in our starting fleets then the pain of catching up to make more is WAY lessened.

They only have one slot anyway, helps justify the free ship.

Oh ALSO thing to take note of: a damaged collector when docked will stay docked until they’re healed. However they can still be shot at since they’re exposed. This means that if you shoot at a collector while it’s docked but never kill it you can completely halt someone’s collection with a single fighter.

They appear to retain this stupid tactic regardless of setting them aggressive.

1 Like