Why basepushing is so strong

aka, how i learned how to play the game

Discussion about the game and its default mods.
OMnom
Posts: 255
Joined: Fri Oct 28, 2016 4:04 am

Post by OMnom » Sat May 27, 2017 4:38 pm

WhoCares wrote: Ok, i had some game and made some people play, not much to say so far exept that it feels odd to almost all players. I'll play along with this playtest and will try to conduct it to retreave pertinent informations for you and any people really involve in balancing to analize.

I'll need few little things if it's not too much to ask, can you expand this modification to behind the veil and green belt. I feel that depending the kind of map the result and feeling would be different.
http://resource.openra.net/maps/21586/ http://resource.openra.net/maps/21585/

Feel free to go crazy. I'd watch out for the nerd ragers who decide to move their MCVs all over the map and build power plants everywhere....on a more serious note, i'd start with observing the A-moves and the pillboxes.

Disregarding all of the side-conversations, do you guys see how (and most importantly, why) this "one" topic can become so messy? Nearly everyone in here has a different version of what they think is a good fix for basepushing. And as Lorry has mentioned in this thread, along with myself and a few others in all of the other threads, barely anyone bothers to provide evidence for what they say. This kind of shortsightedness is contagious because it's so easy to say "look, there's a hole, lets put something in it," but all you're really doing is fucking shit up.
anjew wrote:
But seriously, economy is relevant in most discussions (idk about shockies). RA's economy is quite unpredictable and as such it allows players to build an army AND defences which in turn allows them to turn these into offensive defence with an MCV. Most RTS games punish players for turtling with defences by not allowing them to make a ground force (or at least a lethal ground force) but in RA you can have your cake and eat it too.
One of the to-do's on my list before my schedule got busy was figuring out what changing the scaling and eco would do to the game....if I can find the time, I will make some of these maps.

User avatar
AoAGeneral1
Posts: 597
Joined: Wed Apr 29, 2015 6:11 pm

Post by AoAGeneral1 » Sun May 28, 2017 12:30 am

@Lorrydriver

I stated it hasn't been tested. That is evidence enough that it is not tested in regards to responses of (It will not work). Might wanna read the posts. (All be it Smitty mentioned the durations were tested but curious in how the outcomes came to be.)

My idea for RA is to be close to RA96 much like TD to CNC95 while keeping a balance premise going. Experience came from Grey Goo and 8-bit armies which by the way both suffer from base crawling/base snaking plays.

@Omnom:

As mentioned once again, not been tested. (Exception mentioned by Smitty).

The economy will be a tough fix since alot of the gameplays is based off its productions and values. Expect a lot of tweakings to happen.

OMnom
Posts: 255
Joined: Fri Oct 28, 2016 4:04 am

Post by OMnom » Sun May 28, 2017 10:23 am

AoAGeneral1 wrote: @Lorrydriver

I stated it hasn't been tested. That is evidence enough that it is not tested in regards to responses of (It will not work). Might wanna read the posts. (All be it Smitty mentioned the durations were tested but curious in how the outcomes came to be.)

My idea for RA is to be close to RA96 much like TD to CNC95 while keeping a balance premise going. Experience came from Grey Goo and 8-bit armies which by the way both suffer from base crawling/base snaking plays.

@Omnom:

As mentioned once again, not been tested. (Exception mentioned by Smitty).

The economy will be a tough fix since alot of the gameplays is based off its productions and values. Expect a lot of tweakings to happen.
We literally just said that you need to get testing to support your changes. You can't use uncertainty of something possibly working as your only source of evidence because that is THE only source of evidence anyone here ever uses. Also, repeatedly chirping "it hasn't been tested" is rather ironic seeing as how you never bring up your own tests.

eskimo
Posts: 325
Joined: Tue Jan 17, 2017 9:59 pm

Post by eskimo » Sun May 28, 2017 8:30 pm

From an outside perspective:

I'll just point out that even if some tweeks don't work they're fun to play.

I always opt to play SoS 1.6 atm as it's fun trying other builds out. Also whatever else i come across.

fernoe
Posts: 28
Joined: Sat Sep 24, 2005 10:21 pm

Post by fernoe » Tue May 30, 2017 11:57 am

Ok, i've briefly read though the thread. i was actually testing out some ideas to limit, but not completely kill, basepushing.

One thing i tested was changing the mcv's buildup time to 300 ticks (i think that = 12 seconds). i discoveed that the buildable area around the mcv doesnt work until the buildup animation has finished playing, so you cant build around it for a few seconds. it also affects the speed of the undeploy animation, which makes it harder to move or sell if its being attacked. It makes basepushing much riskier.

User avatar
JuiceBox
Posts: 168
Joined: Thu Jun 30, 2016 12:10 pm
Location: Liverpool

Post by JuiceBox » Tue May 30, 2017 3:17 pm

@fornoe
I like this idea. I mentioned something very similar awhile back in omnoms tab edit thread. While I like the idea of the slower deploy I don't like the idea of a slower pack up.

The slower deploy gives u ample time to react as the defender while still being deadly if undiscovered.

The slower pack up I think would be problematic. In early game flame rushes green rushes and such. Making your opponent undeploy is victory enough by just setting them back in build time. I could imagine this being open to abuse in early game flame drops.

From what I am reading it seems you can't effect one without the other ?
"I love the smell of JuiceBoxes in the morning"
LT. COL. Bill Kilgore
Apocalypse Now

Holloweye
Posts: 100
Joined: Sun Nov 28, 2010 12:21 pm

Post by Holloweye » Tue May 30, 2017 6:29 pm

Not sure if this already been mentioned (so much to read). But I thought about this simple change:
- Disable defensive structure placement around the deployed MCV for like 30 seconds.

It could nerf a basepush a bit. Giving the other player time to react and perhaps send his force there (if he scouted it before hand).

fernoe
Posts: 28
Joined: Sat Sep 24, 2005 10:21 pm

Post by fernoe » Tue May 30, 2017 9:26 pm

JuiceBox wrote: @fornoe
I like this idea. I mentioned something very similar awhile back in omnoms tab edit thread. While I like the idea of the slower deploy I don't like the idea of a slower pack up.

The slower deploy gives u ample time to react as the defender while still being deadly if undiscovered.

The slower pack up I think would be problematic. In early game flame rushes green rushes and such. Making your opponent undeploy is victory enough by just setting them back in build time. I could imagine this being open to abuse in early game flame drops.

From what I am reading it seems you can't effect one without the other ?
On it's own, deploy and undeploy times are the same, but you might be able to change the undeploy animation with clever use of conditions.

OMnom
Posts: 255
Joined: Fri Oct 28, 2016 4:04 am

Post by OMnom » Tue May 30, 2017 11:24 pm

fernoe wrote: Ok, i've briefly read though the thread. i was actually testing out some ideas to limit, but not completely kill, basepushing.

One thing i tested was changing the mcv's buildup time to 300 ticks (i think that = 12 seconds). i discoveed that the buildable area around the mcv doesnt work until the buildup animation has finished playing, so you cant build around it for a few seconds. it also affects the speed of the undeploy animation, which makes it harder to move or sell if its being attacked. It makes basepushing much riskier.
I haven't done this directly, but I did try putting an 8s delay as a BuildRadiusDelay whenever you deployed an MCV, which has a very similar effect to this. I chose 8s as the delay because that is the unboosted build time of a single powerplant, so the pace at the start of the game wouldn't be affected too much. The concept is good, especially with the undeploy/sell animation, but players got very annoyed with this change when they were trying to expand for money.

A bigger problem, however, is that since this change is applied to the MCVs of both sides, it becomes more difficult to uproot someone from a fortified position. This has a lot to do with the vision system, the building queue mechanics, and the strength of static defense, just to name a few. The end result is that a mobile army generally does not have the strength to punch through an encampment (of equal strength) without sheer numbers, taking the time to use arty/v2, or with the help of your own MCV.....in other words, the 12s is a net advantage for an already-deployed MCV. So while this will make mobile armies stronger versus an undeployed basepush, it will also enhance the defender's advantage (which is already very strong). Over the long run, if you only changed this, I'd expect this change to exacerabate the "get there first" mentality and cause games to stalemate more often and/or snowball harder than they are now because of the buffed defender's advantage.

I've been thinking about several ways to offset this buffed defender's advantage, one of which would be giving tanks the "BlocksProjectiles" trait on deploy. I haven't quite figured out exactly how I want to implement that among the 4 tanks though... I've also been considering putting that same trait on buildings, but that opens a very large can of worms.
Holloweye wrote: Not sure if this already been mentioned (so much to read). But I thought about this simple change:
- Disable defensive structure placement around the deployed MCV for like 30 seconds.

It could nerf a basepush a bit. Giving the other player time to react and perhaps send his force there (if he scouted it before hand).
This idea has been brought up in many other threads. As far as I know, there's no way to implement a GrantCondition based off a timer, and there's no condition support for BaseProvider to separate defensive buildings from normal buildings. This is one of the first solutions that I wanted to test, since it would be a great way to cut the power of basepushing without touching MCVs that are used for expanding. One thing about this idea is that it also creates a similar problem as the aforementioned solution, where that 30s creates a gap between the defender and the attacker which benefits the defender, but its not as bad since the attacker can still use buildings for vision, fodder, and reinforcements (and maybe walls).

crlf
Posts: 28
Joined: Tue Jan 10, 2017 9:27 pm

Post by crlf » Wed May 31, 2017 9:03 pm

I think the really difficult part of this issue can be summarised as follows:

- Once an MCV has deployed in an area, that area is almost instantly locked down (probably with a defensive structure); slightly late players are at a disadvantage, even if they have an army, i.e. the "I was here first" problem is undesirable.
- However, it's also a problem that a great way to attack a well-established base is to start building one right under the porch.

Any change has to weigh these two up - it's not even a simple matter of 'who is the defender'.

We want to discourage:

- using newly-built structures to soak damage
- using static defence as disposable offense

They should be exceptional, not the norm.

I would suggest the following measures:

a) Weaken newly-placed buildings

New buildings (including static defence) are deployed at 60% health; having a conyard in build radius grants gradual self-heal. We might need to exclude power plants from this, or change how they scale so they don't lose power till 50%.

Refering back to OMnom's original list, this would take the edge off 2, and also 1, 12, 14. It would favour the long-term defender; in the short term, it makes little-to-no difference.

b) introduce situational cooldown for static defences

Broadly speaking, the objective is to stop defence spamming in offensive outposts while allowing established bases to maintain defences in proportion to their size.

Give each conyard a meter which has as many pips as it has non-production buildings in range (i.e. pp, ref, rd, but not barracks, wf). Each static defence (including walls) costs you a pip. You can only deploy defences where you have pips. Pips slowly fill up like orca ammo. The more buildings inside the build radius, the faster they refill. Undeploy and they're gone.

(Implementation detail: where build radii overlap, non-production buildings grant pips to both; defences take pips from the base with the most pips to spare.)

This would help with 1, 9, 6, 7, 11, 14, 15, 18. It would also strongly discourage the undeploy/redeploy behaviour in 8.

Both of these changes might encourage players to deploy more static defences in likely combat zones during quiet times 'just in case', rather like the AI does, as there's still not much opportunity cost to this, so long as there's cash on hand.

You could also set it up so that the more filled pips you have, the lower the initial health penalty is. Though this also is less of an issue in quiet times.

c) differentiate more between concrete and wood

Give medium, heavy and mammoth tanks more damage vs concrete (the base defence armour type). Weaken allied artillery (and maybe V2s) a bit vs concrete. The aim here is to penalise pure turtle arty+hind+inf builds which these other changes would otherwise buff - they either need longer to chip away where defences exist, or they need to be willing to go in and take losses, or they need to spend wf time on tanks.

I appreciate that a and b are big changes which require code and would change the game a lot!

User avatar
AoAGeneral1
Posts: 597
Joined: Wed Apr 29, 2015 6:11 pm

Post by AoAGeneral1 » Wed May 31, 2017 10:01 pm

crlf wrote: I think the really difficult part of this issue can be summarised as follows:

- Once an MCV has deployed in an area, that area is almost instantly locked down (probably with a defensive structure); slightly late players are at a disadvantage, even if they have an army, i.e. the "I was here first" problem is undesirable.
- However, it's also a problem that a great way to attack a well-established base is to start building one right under the porch.

Any change has to weigh these two up - it's not even a simple matter of 'who is the defender'.

We want to discourage:

- using newly-built structures to soak damage
- using static defence as disposable offense

They should be exceptional, not the norm.
Agree strongly. Prefer to have it not be the 90% way to win games. Exceptional I agree.
crlf wrote: I would suggest the following measures:

a) Weaken newly-placed buildings

New buildings (including static defence) are deployed at 60% health; having a conyard in build radius grants gradual self-heal. We might need to exclude power plants from this, or change how they scale so they don't lose power till 50%.

Refering back to OMnom's original list, this would take the edge off 2, and also 1, 12, 14. It would favour the long-term defender; in the short term, it makes little-to-no difference.
This could possibly work in that situation alone. If someone has made a drop of grenadiers in the base or a flamer drop in the main base you are out of luck. Dropping a defense structure to prevent this with a HP reduction is then instantly killed. I like the handicap idea but the handicap can be placed in the wrong situations too.
crlf wrote: b) introduce situational cooldown for static defences

Broadly speaking, the objective is to stop defence spamming in offensive outposts while allowing established bases to maintain defences in proportion to their size.

Give each conyard a meter which has as many pips as it has non-production buildings in range (i.e. pp, ref, rd, but not barracks, wf). Each static defence (including walls) costs you a pip. You can only deploy defences where you have pips. Pips slowly fill up like orca ammo. The more buildings inside the build radius, the faster they refill. Undeploy and they're gone.

(Implementation detail: where build radii overlap, non-production buildings grant pips to both; defences take pips from the base with the most pips to spare.)

This would help with 1, 9, 6, 7, 11, 14, 15, 18. It would also strongly discourage the undeploy/redeploy behaviour in 8.

Both of these changes might encourage players to deploy more static defences in likely combat zones during quiet times 'just in case', rather like the AI does, as there's still not much opportunity cost to this, so long as there's cash on hand.

You could also set it up so that the more filled pips you have, the lower the initial health penalty is. Though this also is less of an issue in quiet times.
Few questions on this:

When you mention about buildings inside the radius are the buildings giving faster ammo regen to the conyard itself?
Do defense structures count as that ammo regen?
Does this nerf base expansions a bit to much in the sense that you expand and a small scout force happens to spot it. But you can't deploy a defense structure because you have no pips. Would this cause issues?
Moving your MCV inside your own base or the fog of war trick undeploy/redploy cause to much issues with this?
crlf wrote: c) differentiate more between concrete and wood

Give medium, heavy and mammoth tanks more damage vs concrete (the base defence armour type). Weaken allied artillery (and maybe V2s) a bit vs concrete. The aim here is to penalise pure turtle arty+hind+inf builds which these other changes would otherwise buff - they either need longer to chip away where defences exist, or they need to be willing to go in and take losses, or they need to spend wf time on tanks.
I do prefer if tanks can do a little more damage vs structure types as they are lacking in regards. This problem also persists with heavy tanks and mammoth tanks having a slow rate of fire in compared to medium tanks. (Mammoths shoot slower then heavies in one example). If raw damage was to be increased vs structures its possible that could be compensated. Mammoth tanks have the upper hand to split their fire from canon and missiles for maximum damage but requires micro.

A problem I recently found out is you can build both concrete walls and wired/sandbags without even building a power plant. Why?

Change this to where the wired/sandbags can be built but the concrete walls need a war factory before they can be built. It doesn't make much sense why these two can be built hand in hand.
crlf wrote: appreciate that a and b are big changes which require code and would change the game a lot!
Appreciated on the input ^^

fernoe
Posts: 28
Joined: Sat Sep 24, 2005 10:21 pm

Post by fernoe » Thu Jun 01, 2017 3:42 am

OMnom wrote:
fernoe wrote: Ok, i've briefly read though the thread. i was actually testing out some ideas to limit, but not completely kill, basepushing.

One thing i tested was changing the mcv's buildup time to 300 ticks (i think that = 12 seconds). i discoveed that the buildable area around the mcv doesnt work until the buildup animation has finished playing, so you cant build around it for a few seconds. it also affects the speed of the undeploy animation, which makes it harder to move or sell if its being attacked. It makes basepushing much riskier.
I haven't done this directly, but I did try putting an 8s delay as a BuildRadiusDelay whenever you deployed an MCV, which has a very similar effect to this. I chose 8s as the delay because that is the unboosted build time of a single powerplant, so the pace at the start of the game wouldn't be affected too much. The concept is good, especially with the undeploy/sell animation, but players got very annoyed with this change when they were trying to expand for money.

A bigger problem, however, is that since this change is applied to the MCVs of both sides, it becomes more difficult to uproot someone from a fortified position. This has a lot to do with the vision system, the building queue mechanics, and the strength of static defense, just to name a few. The end result is that a mobile army generally does not have the strength to punch through an encampment (of equal strength) without sheer numbers, taking the time to use arty/v2, or with the help of your own MCV.....in other words, the 12s is a net advantage for an already-deployed MCV. So while this will make mobile armies stronger versus an undeployed basepush, it will also enhance the defender's advantage (which is already very strong). Over the long run, if you only changed this, I'd expect this change to exacerabate the "get there first" mentality and cause games to stalemate more often and/or snowball harder than they are now because of the buffed defender's advantage.

I've been thinking about several ways to offset this buffed defender's advantage, one of which would be giving tanks the "BlocksProjectiles" trait on deploy. I haven't quite figured out exactly how I want to implement that among the 4 tanks though... I've also been considering putting that same trait on buildings, but that opens a very large can of worms.
Holloweye wrote: Not sure if this already been mentioned (so much to read). But I thought about this simple change:
- Disable defensive structure placement around the deployed MCV for like 30 seconds.

It could nerf a basepush a bit. Giving the other player time to react and perhaps send his force there (if he scouted it before hand).
This idea has been brought up in many other threads. As far as I know, there's no way to implement a GrantCondition based off a timer, and there's no condition support for BaseProvider to separate defensive buildings from normal buildings. This is one of the first solutions that I wanted to test, since it would be a great way to cut the power of basepushing without touching MCVs that are used for expanding. One thing about this idea is that it also creates a similar problem as the aforementioned solution, where that 30s creates a gap between the defender and the attacker which benefits the defender, but its not as bad since the attacker can still use buildings for vision, fodder, and reinforcements (and maybe walls).
Instead of having the MCV not allow base defense for 30 seconds (which is pretty tricky to code) you could make the buildup animations of base defense take 30 seconds. This would have similar effect on gameplay because base defense structures cannot fire until their buildup animation is complete.

OMnom
Posts: 255
Joined: Fri Oct 28, 2016 4:04 am

Post by OMnom » Thu Jun 01, 2017 8:37 am

fernoe wrote: Instead of having the MCV not allow base defense for 30 seconds (which is pretty tricky to code) you could make the buildup animations of base defense take 30 seconds. This would have similar effect on gameplay because base defense structures cannot fire until their buildup animation is complete.
increasing the build duration of pillboxes to 30s / anything equilvalent isn't a road I want to go down on. If someone else wants to be in charge of this testing, feel free to do so. Someone tried making all defensive structures take 100% longer (i think it was amhol) The immediate impact that this had on the game was extremely drastic; early rushes are near impossible to stop without going 2ref yourself (limiting build order diversity), your economy is in constant danger and is extremely difficult to stabilize, and any setback you run into is amplified. If you changed this, you'd have to change the entire infantry foundation along with it, which doesn't do anything except buff vehicle usage (which are already very weak)...if it comes to this, we might as well just turn the entire game into TD. This is a change is a very large can of worms, one which I am not ready to open.

As much as I hate to admit, the current static defense set up is the only thing stabilizing the early game. Yeah, it could use a larger nerf, but 18 or 20s is the highest we can go to without experiencing destabilizing effects....that would be the equivalent of my $700/ $800 pillbox tests. I'm exploring other options at this moment.

User avatar
AoAGeneral1
Posts: 597
Joined: Wed Apr 29, 2015 6:11 pm

Post by AoAGeneral1 » Thu Jun 01, 2017 9:53 am

Not entirely sure how this causes problems for early infantry attacks if you already have the structures in place to defend this. You can also leave one queued up as many players do already.

30 seconds seems a bit to long though and would prefer something along the lines of 22 seconds. Plus having multiple MCVs reduces the build duration by quite a lot. (IE: current maximum MCV counter decreases tesla coil build times to 15 seconds. Which is insane.)

This actually wouldn't turn into TD from this change itself since MCV expansions are much easier to produce and infantry are stronger in RA. (Funnily enough however infantry have recently become a bit popular in TD.)

OMnom
Posts: 255
Joined: Fri Oct 28, 2016 4:04 am

Post by OMnom » Thu Jun 01, 2017 10:20 am

Not entirely sure how this causes problems for early infantry attacks if you already have the structures in place to defend this. You can also leave one queued up as many players do already.

I'm not going to play the theorycraft game with you right now. Literally, every conversation we have is moot, goes on for 2 pages, and ends with both sides saying "needs more testing," with me being the only one ever coming back with any type of data.

Post Reply