So I am designing a system involving the Human Direct Damage System (Uses Move Scarab to LocationOverWater, which causes said scarab to explode).

I have everything working swimmingly except... the storage system doesn't work quite right simply because I (unless I am mistaken, or have forgotten how to) cannot stack the scarabs.

Anyone have any ideas or theories on how to efficiently and neatly store human scarabs?

Edit: Attachment of current system. [EOS Testing2]
Edit2: Removal of old system, added link to new one.

Post has been edited 2 time(s), last time on May 4 2010, 11:01 pm by TiKels.

The DDS system that most people know is very inefficient. I've made one that uses only one location (if you exclude the unwalkable location and the one that center around the target) and three triggers. Attached is a system I've made myself and tested. As far as I know, it's the best way to use DDS (both for humans and computer). Correct me if I'm wrong.

Your system has a fair chance of moving a scarab that is already "dead" and failing to deal damage. I'm looking for 100% accurate because I'm going to need 100+ over a few seconds. Plus your scarab production is slow anyway.

Puttin' an attatchment on front post.

Through extensive testing, I found that the probability of moving a dead scarab is around 3%. By my standards, that is more than sufficient. I don't understand how you can be picky about a 3% error when you're moving over 100 scarabs but it's your call.

Generating hundreds of scarabs every few seconds will take a lot of space and triggers using regular DDS. In fact, it's much easier to generate a great amount of scarabs using my method; you can make a bigger square (or several using burrowed units in each and moving a location across every square) and put walls with eggs on all four sides of each square. Also note that I did not set the timer to let the reavers fire only once; reducing it further would greatly increase the production of scarabs. With the regular DDS system, you'd need to make 100 little rectangles and 100 reavers (again, if you want 100/second).

Otherwise, did you check the DDS demo available on the DDLB (http://www.staredit.net/files/1618/). I've not had time to examine your map but what's available seems to work perfectly. I'll give you further feedback tomorrow if I find out what's wrong with your map.

I am going to have each scarab actually successfully hitting, not -maybe- hitting, so a flawless system is what is needed (IE: Not your system). I could just string out the scarabs onto 4/8 different rows instead of just one, like in the Human DDS map you posted, so it doesn't clump up majorly. The system I use is dynamic in that I can add in more rows and only have to copy 2 triggers, and also that each box doesn't have a location over it, but just a single burrowed unit. I just simply detect when a scarab is present when I cycle through them. EZ123.

I don't understand why you move scarabs to stack then to storage instead of straight to storage.

Moreover, you will very likely end up using dead scarabs since the "scarab grab" location covers the entire lane; the location does not cover the end of the lanes in the Human DDS system because the scarabs that reach the end will be dead. Also, if you use only one lane, the scarabs will cluster at the beginning and die there; again, you will end up moving dead scarabs.

You said your system does not need a location over each box but neither does the human DDS one. In fact, the only way you can improve on that system would be to eliminate the DDS_Target location, extend DDS_Storage to cover the scarabs and their target, use only one type of target unit and use a timer to replace all the reavers instead of checking if there's a scarab every time (since they attack at precise intervals; that would save a great deal of triggers). I have not tested those alleged improvements but I'm confident they would work. Of course, the system as it is is already sufficient to generate the results you want.

You should also move your line of scarabs to the left of the reavers so that the location grabbing scarabs will grab leftmost ones, which are usually about to explode farther in the future than the others.

I don't think so. All units in SC seem to have random intervals between shots. I know marines do for a fact - if two marines with equal HP and upgrades engage each other, one would always be first to shoot so you might think that marine would always win, but that's not always true (even though it is most of the time).

Reavers in the DDS seem to have varying attack intervals; maybe because they don't spawn at the same angle and some need to turn more than others.

Broodlings have extremely pronounced randomized attack intervals. I did not know that other units are affected by this, and it would be interesting to see which ones do.

I'm well aware that there is a variance but that variance is neglectable. It takes around 3 or 4 loops (with hyper triggers) for every reaver to attack, regardless of initial conditions. Using a timer instead of detecting scarabs can be very useful if you're using hordes of reavers since you would only need half as many triggers (given that there is one trigger to center the "detecting location" and another to detect if there's a scarab for every reaver). I've tried it already with a timer and it works pretty well.

Apparently "Bring one scarab to Location X" on a Reaver holding a scarab (but not firing!) is still true, which (possibly) explains the erratic spawning, where a scarab is ONLY made when they are created facing the exact right direction. Doing testing ATM, will get back to you.

This COULD work, if, human scarabs were actually targeting anything. They are just by default heading towards the coordinate 0,0.

Actually JaFF is right, and it makes a pretty big difference. The way I am doing it, I may take 2x as many triggers, but what is 2x as many when the original is only 8ish. The fact is, if you don't spend the time to make 8 more triggers to PERFECT your system, it's just wasteful. My system works very well, albeit taking more units.

System perfected; If anyone cares I'll post the perfected system.

A perfected DDS system?!? I CARE!!

You're right Tikels. If you use a timer, there's a 20% difference in scarab production. That is, there were between 40 to 60 scarabs present when using a timer and between 50 to 70 when using detection. It's not a big difference but if you really want the "perfect" system, then you have no choice.

As to the detecting scarabs when they're in hangar, I'm certain JaFF is very well aware of this(since he made that human DDS map). If you look at the map, you'll see that the detect location is just big enough not to encompass the entire reaver when centered on the target; if you made the detect location a single tile bigger, the whole system would fail working for the reason you mentioned.

The only thing that is wrong with it is that one of the rows (every now and then) will have a reaver spawn facing downward a few times in a row, lowering his turn-around time to 0 frames, allowing a reaver to shoot QUITE quickly, gamming up the system. This can be avoided by refusing to move the scarabs over if there is a scarab at the start, and instead just removing it entirely. Going to run a stress test later today, see an average percent of failure without the aforementioned prevention system. See how many hits out of 10,000 mess up. Will take 13 minutes and 53 and 1/3 seconds to test, so I don't wanna do it now.

Attachments:

Couldn't EUDs work as DDS if we enable them?

If you wanna force the use of EUD Enabler and render the map unplayable for Macs (and anyone who hasn't heard of EUD Enabler).

Those two statements are very contradictory; you cannot generate that many with only 8. Finding a way to halve triggers, albeit not necessary, would be very elegant.

For someone who is seemingly so obsessed with perfection, your "perfected system" does not really live up to its name. Firstly, you use way too many locations. Secondly, there is absolutely no difference with the "human DDS" one (aside from the extra locations).

Moreover, if you stop moving scarabs when you detect one is already in the lane where you're going, it will greatly reduce the production of scarabs for no reason; that is, I think that scarabs stack and stay idle only if several are moved to the same location during the same trigger loop (which cannot happen since there are as many lanes as scarab-moving triggers). I'm almost certain that the system as it is will never produce clogged scarabs at the beginning of the lanes.

Here's a truly perfected system. It uses only 3 locations and 1 trigger per reaver. Above all, its scarab production ranges between 70 and 90 (the regular DDS ranges between 60 and 80) for 4 reavers.

No. I've had them blow up at the start of my lanes, if you watch my map long enough it will happen a few times. If u place (random number) 7 scarabs over 7 trigger cycles it will jam it up. Plus you could always keep a "Reserve lane" or two where extra scarabs could go, if you don't like stunting your spawn rate.

Alright, you beat me on efficiency in terms of locations and triggers.

... I could actually do it with two locations, but, hell, who cares.

In my system there are 100-120 scarabs in the system at all times.

There are 8 reavers in your map. That means there would be 50-60 with 4. As I've stated it before, regular DDS yields 50 to 70 (I made a mistake in my previous statement) and the one I posted last was 70 to 90 (no mistake there). The reason for the lower number in the regular DDS is that, the reavers have to fire in the order in which they are selected by the detecting location (since the location centers on the first, checks if there are any scarabs then goes to the next). Whereas in mine, the order of the reavers' attack has no impact whatsoever since there's one huge location encompassing all targets.

You're right about the jamming scarabs: just add as many lanes as you want until they stop clogging altogether; it does not require any additional triggers and you only need to stretch the storage location a bit. Obviously, you will probably be fine with one or two additional lanes since the reaver's attack speed is restricted to a relatively small interval of finite values. Removing scarabs to prevent clogging is unacceptable if you're aiming for the perfect system (gob forbid we should cut down production even by a single percent ).