Do you think late T4 stage on land/navy maps is too short due to T3 arty or game enders?

maudlin27

Not for the games I play in (mostly 3v3 mapgens)

I can't remember the last 3v3 TMM game I had which featured a game-ender, possibly it never has. So no, I don't think it's too short due to them.
T3 arti gets seen occasionally, but the game is usually over by then (as if one team can build a t3 arti without losing ground significantly then they were likely already well ahead), so I don't think it ends the game too early (and 1 T3 arti can be defended against anyway if the game is close)

Nuggets

I'll comment here aswell: Yes, 100%. As mentioned numerous times by many people; game ender such as para and yolo is too easy and most often too hard to punish, unless its a 3v3 or less where they are barely even made.
Setons for example heavily favors game enders over any sort of land aggression and id say many other maps too. The problem is that an attack / invasion is harder, takes more time and is more risky, while with game ender you make defenses and then lean back until its done.

@maudlin27 If a game ender (or even t3 arty, but mostly enders) isnt seen in a game because the game is too short, its kind of an irrelevant comment, no? No offense.

My proposal would be to make ALL game enders more expensive and give mavor and scathis a buff to make them actual "game enders". T3 arty would probably need an adjustment based on that aswell, but i cant comment on that...

Skrat

I consider the current balance of the late game to be good and not requiring changes. It all depends on the type of maps and the personal preferences of the players.

Some people like to play a full game with units, whilst others prefer to rush their economy and build a game-ender; there’s a suitable map or game mode for every player (hello Dual Gap, Astro, Ladder, etc.)

I can guess which replay prompted the Nuggets to start this discussion.

An attack/invasion on Seton is very powerful and more effective than a game-ender. It’s simply a matter of timing. If you’ve won your naval side at the 15-, 20-, 25- or 30-minute mark, it’s more advantageous for you to start an invasion. Because your economy isn’t yet strong enough for a game-ender, and there isn’t much reclaim left in the water. However, if you’ve won your side at 40+ minutes, there’s 100–150k reclaim left in the water, and you spend 5 minutes building T1 factories with a full mass storage instead of investing that massive amount in a game Ender - whilst your opponents in a similar situation are building a game Ender and you end up losing - then this isn’t a balance issue and isn’t a reason to make the game Ender more expensive.

There are players who want to spend 1.5 hours playing solely with units, without using nukes, T3 artillery or game enders. I don’t think that’s a reason to remove all ‘game-ending’ structures from the game. Just as it’s not a reason to completely overhaul the game’s economy for the sake of those who want to play only T1 stage up to 20 minutes.

I also believe there is no need to change the balance between the game-enders. Each of them is effective in a specific game situation. Perhaps it is only Scathis that needs to be improved or made cheaper.

Skrat

Game-enders are generally almost never built, even in large 6v6 or 7v7 team matches. On Mapgen, even when the total mass income per team is 1,500–2,000, players prefer to build EXP or T3 artillery. If we make game-enders even more expensive, they’ll only continue to be built on dual gaps. Let’s not remove 5 buildings from the game.

Nuggets

The strategy to "rush" enders will still be there, just take longer. Thats the entire point. Give invasions / pushes more of a chance.
Saying pushes / invasions are generally stronger is just objectively wrong. This is not my opinion this is a fact. I never make game ender not because its the better move, but because i find the alternative boring af. Sure thats my opinion and thats why i dont want to make this the only choice, but it does need adjustment.

Yes, there are maps and big teamgames where they are not made, but often thats not the optimal move. And thats is what im trying to adjust: shifting the balamce so that ender is no longer the optimal move, but rather one of 2 choices that are equally balanced

(Not to mention i find the concept of "rushing" a "game ender" a bit ridiculous)

Skrat

@Nuggets In your opinion, what should the timing be for a game ender?

In my view, there are five stages to the game: T1, T2, T3, T3+exp, and Game-enders. In my experience, stages T1 and T2 last an average of 5–7 minutes each (for naval units, it’s a little more). Stage T3 and stage T3 + exp each last an average of 10–15 minutes. And after 40 minutes of play, the ‘game-enders’ stage begins (if you choose the most effective way to win, and don’t refuse game-enders on purpose)

If you wish to extend the duration of stage T3, I would only agree to this on condition that stages T1 and T2 are also changed to increase their duration. I don’t think it’s right to spend 80% of the game solely on stage T3

Also, I don’t understand to what extent you want to increase the cost of game-enders, so that with 150k mass on the reclaim after a side’s win on the seton, it would be not profitable to build a game-ender? It may be worth increasing time to eating reclaim and it will make it more difficult to reinvest a lot of mass in game ender so quickly. But this is too big game-changer that will affect all stages and requires a separate discussion.

FtXCommando

The conversation is pointless without deciding upon a baseline of maps to center the discussion around. If we operate by the metric of matchmaker pools, which is how FAF historically determines the framework for balance, then game ender balance is fine in my mind. I don’t really see many 20x20 maps in the 2v2 or 4v4 pool that make t4 gameplay unfeasible or too short. Likewise, 3v3 tmm doesn’t generate maps where game ender before min 40-50 makes much of any sense.

Now if we instead operate on the 6v6 crazy mass map gens that often get hosted on weekend, then yes game enders come online super quick if people are playing competitively, but I just disagree on basing something like game pace balance off of these lobbies.

I personally think the only problematic game ender units are

Paragon and Yolo as Nomander mentioned. Yolo should not be the cheapest game ender and paragon needs a large nerf to resource generation
Nukes, they are toxic to the game with how they can swing games due to a lapse in base scouting for 90 seconds. It should require utterly absurd misplay to have something like a main base nuked. I’m talking like not building your SMD until an SML is 75% done.

I think it needs to be easier to dump the resources into an SMD than it currently is, whether that’s making SMDs have more bp or adjust the ratio so engies provide more of a relative boost. The fact a nuke + sml costs close to a mainline t4/half a t3 arty is crazy for the value it provides. The sheer existence of it can often morph entire game plans for the next 5 minutes to purely revolve around it for both teams.

T3 arties are fine, maybe like a 10% cost increase but I wouldn’t do anything more than that.

Novax should just be adjusted to be a high alpha weapon so engineers can properly deal with it by rebuilding behind things it kills.

maudlin27

@Nuggets said:

@maudlin27 If a game ender (or even t3 arty, but mostly enders) isnt seen in a game because the game is too short, its kind of an irrelevant comment, no? No offense.

It's very relevant. If a game-ender doesn't feature in games that get to the T4 stage, then the T4 stage on such maps was self-evidently not "too short". It'd potentially also be an indication that the T4 phase could be too long on such maps. It makes no sense to focus solely on the tiny % of games where game-enders feature and decide only based on those games whether the T4 phase is too long. Game-enders serve a valuable role in preventing very long protracted games, and so account should be taken not just of the games they did end, but the games they didn't (that reached T4), and whether that range of time is considered desirable (while also factoring in that even if it was decided that games should be able to go on for many hours like in LOUD, nerfing them into the ground would also decrease the variety of units and strategies, which I'd see as a negative).

The strategy to "rush" enders will still be there, just take longer. Thats the entire point. Give invasions / pushes more of a chance.
Saying pushes / invasions are generally stronger is just objectively wrong. This is not my opinion this is a fact.

It's very much an opinion whether T4 pushes are stronger than game-enders. Something that is fact would be saying something like 'In 4v4 tournament games featuring a large cash prize and top rated FAF players, there were several games where one team focused on a game-ender, their opposing team focused on a T4 push, and the team focusing on the T4 push won' (while an opinion would then be if I said I think that shows that T4 pushes can be more effective than game-enders).

Nuggets

FTX automatically jumps to the insane mex mapgen maps when we also have setons as an example for moderate mex amount. But i do agree with ftx that we cant just put a number regarding match length down because obviously it cant be applied to all maps and sizes.

@maudlin27

It's very relevant. If a game-ender doesn't feature in games that get to the T4 stage, then the T4 stage on such maps was self-evidently not "too short". It'd potentially also be an indication that the T4 phase could be too long on such maps.

No? Sure maybe it can mean that, but it can also not mean that. Just because a setons for example ended min 10 doesnt mean anything. It means that that match was a shit example and irrelevant to this.

It's very much an opinion whether T4 pushes are stronger than game-enders. Something that is fact would be saying something like 'In 4v4 tournament games featuring a large cash prize and top rated FAF players, there were several games where one team focused on a game-ender, their opposing team focused on a T4 push, and the team focusing on the T4 push won' (while an opinion would then be if I said I think that shows that T4 pushes can be more effective than game-enders).

No. Obviously you cant generalize "game enders are better than t4 pushes" because there is a difference between t4 pushes and the maps themselves.
And VERY often people on purpose go for a push even though they know game ender is the better move, just because its a lame strategy. But not everyone agrees that its lame, thats why some games are still ended by a "too fast" game ender such as para or yolo.
I dont know how exactly to put this into words, but maybe someone else can explain this better than me. But mostly it always feels like most people cant be bothered to explain anymore because of this shitty state we have been in for years now with no change in sight.

KnownSniper

game ender stage arrives too early. the entire design is supposed to end a game that is locked and nobody can make any progress, but you often see game enders being build in a completely dynamic game. Game enders should not be able to win against a team that's actively trying to play the game instead of sitting afk in their bases watching the progress bar of the game ender go up slowly and then win when it's done.

Skrat

The 220k of mass that you need to spend on average on game ender is a lot. That's a lot. If your opponents in a dynamic game were able to invest a mass equal to 9 GC or 25 Battleship and you did not punish them for this, then this is not a balance problem.

Again, a lot depends on the maps. The larger the map, the lower the exp efficiency due to the fact that it takes you a long time for exp to walk from one side of the map to the other side or does it take a long time to prepare and defend bp closer to enemy bases. And I don't see anything illogical or wrong with that.

Nuggets

You cant simply compare the mass cost. Thats the exact problem.....

That mass that you invest in units are sometimes factories, and there is builduptime, traveltime and so on. Its not simply "your para is done, my 9 gcs are done, lets move in and see what happens".

The fact that i have to mention this simply means you never noticed this or were in this position, because i assume in that situation you would just make a game ender. Slight difference in apm and effort required!

Skrat

@Nuggets I know what you're talking about. Maybe we don't fully understand each other because of the translation. I mean, what if builduptime and traveltime and so on not allowed to punish your opponent for building a game ender, then this game can no longer be considered dynamic. Which means building a game ender is an absolutely normal way to complete the game.

This does not happen on 10x10 maps or maps with a small amount of resources. This doesn't happen in a 1x1, 2x2, or 3x3 game. But if you are playing 20x20km slopgen, where there is a huge amount of resources, or Seton, where in late game navy gives a huge amount of reclaim that is building a game ender is an absolutely normal way to complete the game.

Not everyone wants to play one game for 1.5 hours, building 50 exp each, as if it were a t1 unit

FtXCommando

It’s just a paragon/yolo problem on sentons where paragon pays itself off in 30-60 seconds and yolo has the brutal reality of smds where increasing bp isn’t a particularly great option to deal with it (unlike bp on shields for mavor and scathis or outbuilding the destruction of pgens) and those smds can still get edge nuked.

KnownSniper

@Skrat said:

Not everyone wants to play one game for 1.5 hours, building 50 exp each, as if it were a t1 unit

that's not what's being talked about at all. I have seen tons of games where the optimal move is to make a game ender somewhere between the 30 and 40 min mark. The most classic example of this is 2 opposing sentons ponds winning their navies but 1 guy goes for a proxy while the other goes for a game ender (both starting at min 35 so its still "early" on in the game, not 1,5 hours in). It is way too easy to have a good shot at winning the game while doing nothing while watching your game ender slowly tick up compared to putting in 100x the effort in a proper invasion just to lose to a game ender.

Also it is true like ftx said that a paragon and yolo are the 2 main contributors to this imbalance. It used to just be a paragon that's op but since the yolo bugs got fixed it's now way too strong.

@Skrat said:

This does not happen on 10x10 maps or maps with a small amount of resources. This doesn't happen in a 1x1, 2x2, or 3x3 game. But if you are playing 20x20km slopgen, where there is a huge amount of resources, or Seton, where in late game navy gives a huge amount of reclaim that is building a game ender is an absolutely normal way to complete the game.

yea it doesn't happen on 10x10 maps, that's because t3 arties are the game enders on those maps which you do often see.

Aside from that i don't understand your argument about nto seeing it happen in 1v1 to 3v3's. Does that matter at all? We saw 0 game enders being build in LOTS so it doesn't matter that they're breaking teamgames!

If anything it's the opposite that if they were made more often in lower player count games it's an incentive not to nerf them because it would ruin the unit for those type of games, but as it stands you basically never see them being build anyway so that is of no concern.

Skrat

@KnownSniper Seton is very different from other maps by having 2 large water pools. I'm not sure if we should build a balance based on this map alone.

On land maps, if you build exp instead of game enders in an equal situation, you will win in most scenarios because your exps can safely reach and kill enemy bases. On most water maps, a navy victory will be enough, as the t3 range of navy units will be able to destroy all bases.

And only on Setons and similar mapgens is the mid too small to let exp through it (since they will be destroyed by the enemy navy from another side), but at the same time the air base is far enough away to be out of the firing range of navy units. Plus, air base has enough energy, bp, smd, and a convenient place to build a game ender. The choice of game enders is better because of the features of the map, the location of the bases, and the terrain. And not because game enders are too unbalanced.

Therefore, you are unlikely to see game enders on maps like Maridia, Metir, Skadi etc. Even though there are more resources on these maps and the games also last 30-40 minutes or more.

If you don't build game enders in a situation where you've won the navy side on Setons and you have a good eco and 100k reclaim, then I don't understand in what situations it is necessary to build game enders at all? Only on Gap of Rohan, where there is only 1 gap and it is impossible to break through it with equal play?

Nomander

I got some evidence in the discord discussion after throwing out a 10% cost nerf idea and saying this:

I think I was just throwing out a number [10% cost nerf] to reason about if I was to look at it from nugget's/farm's perspective, but it came off a bit too much like a decision on a change. It was kind of a random number without thought behind it. Personally I'm still not convinced of such a change because like you [maudlin] say there's a lot of different maps that have to be balanced for and maybe setons is just an unbalanced map for this specific scenario, which is unfortunate for its legacy.
More specifically I'm not convinced because:

Nuggets/farms have the argument that invasion is too slow so a cost nerf is needed.

their evidence is "I've seen games where it's optimal but they didn't do it"

no replays where it's optimal but wasn't done

no replays where this optimal move was done and the enemy invasion was equal mass but simply couldn't win, particularly when not vs para/yolo

they themselves use house rules to not have any recent experience using or playing against the strategy

they both say its true on mapgen as well but there are no replays and no explanations why. I also checked nuggets' replays and saw that he often plays drawn out/setons-like naval mapgens and such types of maps that aren't relatable to the average mapgen player, so taking that into account lowers my confidence.

ftx says its purely a para/yolo problem, so a cost nerf is not needed.

All of these I guess you could say they're too low-rated players just going off experience (quite harsh ik but like there might be some errors to look out for, like maybe opps are too bad at defending):

skrat and spcr (mlgtastic) argue that the ender's cost gets it killed by land

sainse argues t3 arty would be unbalanced and claims that the timing just isn't possible/doesn't happen as early as nuggets says it is

maudlin claims the timing isn't an issue

Nuggets and farms responded that searching replays is too tedious compared to the chance of a satisfying balance conclusion, but farms still gave a replay.

From @maudlin27; Rainbow cup game where T4 land push wins vs ender:

FYI this was the game I recalled (T4 land push vs game-ender):

https://www.youtube.com/watch?v=Tdv1ZoqBvbc

24329811

20x15km map, has AI enemy teams in the corners which complicates things, but has the benefit of being a cash prize tournament so you'd expect teams are going with the strategy they think is best.

Using in-game timestamps (rather than video):

at 29m40 teams are roughly equal on eco (1.1k mass), and they largely remain similar after this for a while (bottom team fluctuates between 1.2k and 1.0k)

The mavor is started by bottom team at some point before 34m (at 34m17 it's 32% done); the top team briefly started a paragon before abandoning it and focusing on building GCs and other units

Mavor completes at 39m

Top team pushes with GCs, bottom team manages to land a nuke on a couple of the GCs (and later the AI also nukes one of top team's GCs). was very close but top-team's push with the GCs wins.

From @knownsniper; public/casual high-rated setons where he plays without eco bar (!!giant blunder!!)

anyway just for you this is an example #27433457

due to a disconnect the replay stops at min 38

https://puu.sh/KQj23/ea24a46581.png

in a completely even position (income is the same total mass generated is the same, also 1 min after this time nory wins top navy so both navies are equal)

but there's this to watch the end https://www.twitch.tv/videos/2822304906?t=05h11m32s

Here's @sainse's local replay that doesn't end early.

My take on the rainbowcup replay is that units showed to be better than Mavor. Nukes hindered the land push a lot but then air came to save the push, although everything was still quite slow. Thankfully Mavor is weak compared to Para/Yolo, particuarly vs units, so it didn't deal complete game-ending damage in the time it took for the land push to win.
It might've been a bad emergency defensive situation for the Mavor because there was Fatboy spam instead of other T4s. If the fatboys were swapped out fully in cost then maybe the mavor would've stayed alive with most of the bases destroyed, and the game would've continued for a while, which is a fair situation imo.
If Para/Yolo were built then they defeat the push easily and the north team would've lost because they had no follow-up with arty or ender (which can potentially defeat a yolo that spent its time nuking units, but they can't win long-term against para's defensive shield and smd capabilities).

For the setons replay I think:

Popiska and farms didn't cooperate which led to an ineffective strategy for the team as a whole. They should've either gone farms 2-3 Ahwassa + Popiska 600 ASF or farms should've contributed mass to Paragon. I think the Paragon would've lost but the air attack would've killed the Yolona at minimum and secured future air domination by killing ASF and killing all bp that could've been making SAMs, which allows a followup with more air or a land invasion that skips all the flak setup.
Farms delayed his invasion by 3 and a half minutes because he didn't manage his engineers, factories, and transports correctly. This was because he was focused on irrelevant things as seen in the stream recording, doesn't use ferry orders for transports, and had no eco bar while playing. At 5:13:43 in recording or 41:46 in replay he checked his mass bar and he floated 51k mass and overflowed +145 out of 474 income. It was safe to drop engis for invasion at min 38, Yolona started min 41, and nuke landed min 45.
It was quite lategame so doing an all-in with a land invasion at that timing was already more risky than usual invasion timings.

What I'm convinced in right now is that paragon output should be nerfed and Yolona cost should be shifted from energy to 225k mass to stop air players having a ridiculous 37k/63k mass advantage (17% cheaper than mavor, 25% cheaper than paragon) compared to other enders when turning off their air grid and going into Yolona. 225k mass with lower energy requirements should put it at a similar timing as before for ender rush eco slot build orders (no air grid to pause). It could be slightly lower at 210k mass since it has to eat through the game's accumulated SMDs, but still it needs a significant shift into costing more mass instead of energy.

A work-in-progress opinion is changing Yolona to cost more time when nuking units. A simple answer would be 1.5x missile BT, 1.5x missile HP, and 1.5x minimum reload time, but with how extremely powerful nukes are there could be changes to SMDs from there, so the Yolona change has to wait for whatever happens in that area.

waffelzNoob

In my experience, no. Building a game ender in a dynamic game doesn't randomly result in a victory for the team building a game ender, because the game is dynamic. The team that invested in units destroys bases and gets more map control, which in turn presents opportunities to snipe the game ender. Here's one of the most fun and memorable mapgen games I've played where exactly this is demonstrated, featuring basically every experimental in the game, nukes, novaxes, t3 artilleries, and the game still going on for 20 more minutes until a mavor is built that almost immediately dies to a snipe that wouldn't have been possible without farms's land push: https://www.youtube.com/watch?v=H6rfgLVvZWs (Almost 4 years old, but I don't think interactions changed too much).
This is more difficult on Seton's for example, but if one side won navy you could go through there, especially if they are doing an invasion. Typically a game ender is funded by the entire team, and so to take it down, your team should cooperate fully aswell.

It's also a bit of an oversimplification to say "just watching the progress bar go up" because in the meantime they have to manage defense with much less mass available. The team that built units can also react by spreading out their own infrastructure so that the game ender isn't as effective and buy more time for their units to do serious damage. And yeah, maybe sometimes building a game ender with 200k reclaim is more effective than setting up an entire base from scratch on the opponent's side of the map. Are we going to change balance because of this one-off phenomenon on Seton's?

KnownSniper

@waffelzNoob do you think this about every game ender? Imo mavor/scathis don't really pose an issue but para/yolo are massive problems because aside from being good offensively they're also super good defensively.

Also compared to 4 years ago yolo's were massively buffed to the point were u used to be able to counter them with 6-7 smd's but now its 10-11 (someone correct me if i'm wrong but i keep hearing this from others) because it's no longer as buggy were it doesn't want to launch and/or engi assist is better.

Also if it was just sentons i would agree with you, but in the last year there have been quite some instances with different types of mapgens which let to the same situation as you see on sentons.

KnownSniper

@waffelzNoob said:

In my experience, no

alt text

gotcha!!!