Please upgrade your browser for the best possible experience.

Chrome Firefox Internet Explorer
×

Galactic Starfighter (GSF) Suggestions

STAR WARS: The Old Republic > English > Galactic Starfighter
Galactic Starfighter (GSF) Suggestions

Wildtee's Avatar


Wildtee
02.13.2020 , 12:28 AM | #41
Quote: Originally Posted by sharpenedstick View Post
I'd like to be embarrassed for you, but I know this is literally the best you can do, and you spent all day at the keyboard trying to come up with it. So instead, I'll just say "good job, never give up your dreams!"
Eh, you tried. I'd give this a 2. The best part is accusing someone of being at their keyboard all day when you've been trying to forum pvp in this post for over a day. It's adorable.

Verain's Avatar


Verain
02.13.2020 , 12:46 AM | #42
Quote: Originally Posted by sharpenedstick View Post
Groups produce imbalanced queues.
Ok two things. First, honestly, citation needed. I see tons of extremely imbalanced games in solo queue. Some of them are imbalanced just because I am in them, assuredly a common enough thing for many veteran pilots. Others are imbalanced just because that's how the cookie crumbled. By contrast, there are a tons of really close games when queued as a group- often because another group will be placed as our opponent. You probably don't know that, being that you don't know how to group, so you are already less likely to even be in the game that has less need for solo queue fillbodies.

So I don't think this is even true. If is is true, it's not noticeably true.


Quote:
Imbalanced queues are bad.
Now, moving on to the meat of this- are imbalanced queues even bad? For this, I'm assuming that the first statement- which there's no evidence of- is actually true. So with that hypothetical out of the way....

It is good that the weaker teams lose that is literally the point of competition

The game gives you the ability to team up and be the stronger team. The more people do this, the better games everyone has. The fewer people do this, the worse games everyone has. A good game is not determined solely by the scoreboard, in any event.

Quote:
I want to pwn noobs. So groups good.
You team up to "pwn" everyone. Sometimes you lose, usually you win. All the games are fun. You learn to support your teammates in tough games, figure out how to pair your strengths and cover your weaknesses. If the enemy can't put up a fight, you push them back to their spawn and farm them efficiently, which is not a competitive game, but it's still great to shoot them with blasters.

Grouping is the intended and best way to play this game. We have a discord for finding groups, the GSF Groupfinder Discord ( https://discord.gg/f2VvPB5 ). It took awhile, but it seems like it is actually serving its function of helping players who want to group, find groups.

Quote:
Error to compute. We can hear the gears grinding all the way from here.
You are just mad because all of these things are true:
1)- You can't beat teamwork.
2)- You aren't willing to find a team.
3)- You want the game redesigned to exclude those who are willing to go through the effort of playing it as intended, to cater to your more casual and limited playstyle.

(1) is ubiquitous and (2) is common. What makes you so willing to try to get the rules of the game changed to your benefit and the detriment of other players is (3). That's the part that is so frustrating to so many of us, because it just wastes so much time.

Quote:
I don't know how to multiquote, so I'll just say I've seen Drak's stream, his youtube and him and his stomp squad in game.
Great!

Quote:
Yeah, sure, they don't want 50 to 4 wins. That's why they take them constantly.
I mean, I want the 50 to 4 wins. I want all the wins. Sure, the really excellent premade matches are excellent, but they are all great.

Quote:
It's like they'd love for nothing but 49 to 50s, but why can't they get them?
Scoreboards don't make good games. But we do have plenty of very close and excellent games, and you'll find plenty of them on Drako's youtube channel. So no, we don't have a problem getting good games, or getting close games. But we certainly get stomps in between those, and hey, those are great too!


Quote:
How is piling up all the good players on one team, in coordination and voice chat (though they mostly just shoot the breeze now, because, you know, stomps)
Wait, now we're being called out for having a good time with friends in voice.

Oh, the horror!
"The most despicable person on the GSF forum."

Verain's Avatar


Verain
02.13.2020 , 01:33 AM | #43
Quote: Originally Posted by DakhathKilrathi View Post
Would better matchmaking be nice? Absolutely. But removing premades isn't going to make the matchmaker any better.
While he's dodged this inartfully, the fact is, matchmaker has real difficulty making good matches (with or without premades obviously, exactly as you say), and while I think pretty much everyone is in agreement on this, maybe we need to make more noise about it. It has been quite awhile since we saw the last changes to matchmaker, and while the last two big changes (crossfaction and the combination of "matchmaking based on legacy" plus "groups of N are treated as N copies of the toughest matchmade player") assuredly helped, is anyone thinking that we're anywhere close to peak matchmaker?

When we play as a group, and another group is on, we definitely face them more than half the time. How often should it be, though, that we end up on the same team as them? 10%? That seems fine. But it feels more, like at least 30%.
Many of the players who walk into these threads and ask for "balanced games" do so with the assumption that matchmaker works great as long as there aren't those pesky groups messing everything up. They're wrong, of course, but shouldn't matchmaker be a decent bit better than it is? When I queue solo, no premades are required for matches to be decided before the game starts, and I'm sure you've seen the same. This should happen, of course, but should it happen as much as it does?

Like many GSF things, it's a difficult thing to bring up, because it attracts so many useless comments. Much like how a decent number of people would like a small nerf to proton lockon time, but if you make a thread about it everyone comes in and demands gunship nerfs or that strikes be deleted or whatever.

I don't know what matchmaker does in the background, but I would like the devs to add "have it do a little bit more" to the list of "eventual GSF things". I know nobody can afford to screw it up, but it just seems like something that could get a bit cleverer, and really needs to.
"The most despicable person on the GSF forum."

sharpenedstick's Avatar


sharpenedstick
02.13.2020 , 03:36 AM | #44
Quote: Originally Posted by DakhathKilrathi View Post
I

also who cares if they want to win who doesn't like winning what kind of nonsense is that
You've already conceded the central point with that line. What more is there to talk about?

Either you're for more balanced queues, or you're not.

DakhathKilrathi's Avatar


DakhathKilrathi
02.13.2020 , 05:36 AM | #45
Quote: Originally Posted by sharpenedstick View Post
You've already conceded the central point with that line. What more is there to talk about?

Either you're for more balanced queues, or you're not.
uh huh

Quote: Originally Posted by DakhathKilrathi View Post
I'm going to repeat the very last point in my previous post, since you didn't (and haven't) address(ed) it:

GSF is a group game. You decide whether or not you want to choose who is in your group. You don't have to, but you will be in a group regardless. Matchmaking is essentially RNG, so even if what you are suggesting was reasonable (and it's not), it wouldn't work.


To the second post: The idea that games played or components mastered means anything at all about a player's skill or ability is, frankly, absurd. As mentioned previously. The game already does what you're suggesting, though probably in a way that's better implemented. It still doesn't work because those are silly metrics.

Quote:
I am hopeful that those of us as a player base can come up with constructive ideas on how to better one aspect of our enjoyment that we derive from SWTOR, which is Galactic Starfighter!"
We have suggested an idea for that, that we as players can implement: we can form groups! It's pretty easy, and makes up for bad matchmaking.

Notable: I say this as a dedicated and stubborn solo player. I wish solo was less frustrating, but I don't blame the groups for this problem.
tell me again how i don't want more balanced games

but like

actually have a way to back it up when everything i've said shows clearly that i do, in fact, want to see more balanced games

in other words: actually respond to the points i have made instead of just attacking me as a player

phalczen's Avatar


phalczen
02.13.2020 , 11:34 AM | #46
Quote: Originally Posted by Verain View Post
I don't know what matchmaker does in the background, but I would like the devs to add "have it do a little bit more" to the list of "eventual GSF things". I know nobody can afford to screw it up, but it just seems like something that could get a bit cleverer, and really needs to.
Should matchmaker take into account "skill," "experience," both, or neither?

There was a brief discussion somewhere recently about the merits and shortcomings of win percentage as a general, albeit imperfect, measure of "skill." Accuracy scores could also be a general measure of "skill," and is far less dependent on team dynamics. Can (kill+assist)-to-death ratios, again imperfect, also be used (i.e. a person who can get hits in but is reckless and dies a lot is less skilled than the player who can get hits in AND avoid being downed)? Can DOM medals be used (wins absorb DOM performance, but accuracy and K-D ratio may be less important or telling)?

If any or all of the above four examples are used as indices of "skill," should they be weighted equally, or selectively? Should they be normalized or corrected for the relative balance of DOM v TDM across a pilot's legacy?

I think using "requisition spent on ships on your bar" serves as a somewhat reasonable correction factor to account for a veteran pilot playing on brand spanking new stock starfighters. Should this be weighed equally with the skill factor?

And if this ends up being translated into some numerical score, should each member of a premade be counted individually in determining team makeup pre-pop invite, or should the system remain as is counting the highest member of the group for all four members?
If you think I've made a good contribution with this post, I kindly ask that you use my Refer a Friend link! Here is more information about the program.

Drakkolich's Avatar


Drakkolich
02.13.2020 , 12:20 PM | #47
Quote: Originally Posted by phalczen View Post
Should matchmaker take into account "skill," "experience," both, or neither?

There was a brief discussion somewhere recently about the merits and shortcomings of win percentage as a general, albeit imperfect, measure of "skill." Accuracy scores could also be a general measure of "skill," and is far less dependent on team dynamics. Can (kill+assist)-to-death ratios, again imperfect, also be used (i.e. a person who can get hits in but is reckless and dies a lot is less skilled than the player who can get hits in AND avoid being downed)? Can DOM medals be used (wins absorb DOM performance, but accuracy and K-D ratio may be less important or telling)?

If any or all of the above four examples are used as indices of "skill," should they be weighted equally, or selectively? Should they be normalized or corrected for the relative balance of DOM v TDM across a pilot's legacy?

I think using "requisition spent on ships on your bar" serves as a somewhat reasonable correction factor to account for a veteran pilot playing on brand spanking new stock starfighters. Should this be weighed equally with the skill factor?

And if this ends up being translated into some numerical score, should each member of a premade be counted individually in determining team makeup pre-pop invite, or should the system remain as is counting the highest member of the group for all four members?
Now this is how you start a conversation about the right thing!

Alright so let's unpack a little, you ask if it should use "Skill" or "Experience".

Well it's already using Experience however I'm pretty sure there is currently a cap on it and that needs to be much higher if there is one. It often feels like anyone with atleast 1000 games played is being counted equally by the matchmaker and that just isn't high enough in my opinion to just experience, we have players that have over 15000 games played in the gaming pool.

Next up Skill, I believe we do need to use some metric for Skill, however it needs to be carefully implemented so that players can't abuse how the matchmaker sees them. This is why I'm so in favor of only using Win% because at the end of the day it's the only metric you can't abuse.


I'm gonna throw a few examples at you on how some of the metrics you proposed could be very easily abused.

Let's start with Accuracy, this one is very popular, many people think we could simply add accuracy to the matchmaker to judge players skill. Now what happens when a player wants to abuse that, he simply spends entire matches empty full magazines of rapid fire lasers into mid air, now even though he might have a high win% his accuracy is like 1% and his overall matchmaking rating is going to go down leading to him being able to mess with the system.

Let's look at kill death ratio, you have the same problem as accuracy any game in which the player is winning, they can just throw them selves at a wall repeatedly simply to lower their matchmaking rating.

Now onto Medals in Domination, this one is one of the worst in my opinion. What this one does is reward players for winning by doing the least objectives possible. If I win and get 17 medals and you win and get 3, my matchmaking rating goes up by more then yours. Meaning in the future you'll get better teammates then I will simply because I contributed more in my matches. This is the exact kind of thing we REALLY want to avoid.

The same idea happens with Kills+Assists, If we're winning a game and one person simply decides to stop attacking players they'll get an advantage in the matchmaker.


This is why I weight Win% so importantly above all the stats at the end of the day it's the only statistic you can't "game". You either try as hard as you can to win the games or you don't. If you're losing games on purpose to try to get better teams to win games down the road well that just doesn't make any logical sense.


Last up, you talked about how the numerical score would translate to every player on the premade or if everyone was adjusted to the highest player. I'm very torn on this subject, while I do like the system in place that uses the highest players "matchmaking rating" to determine how to make teams, I also really don't like how it affects relationships of players that have very different skills levels.

For example, I have a few friends that aren't great at the game but enjoy playing none the less, if they come group with me, my ridiculous matchmaking rating puts them at a huge disadvantage because the game considers them way higher then they are. It forces them to choose wether they want a better chance at actually winning matches and playing with players at their level or playing with a friend.

Now having said that, I believe in the long run, matching up to the highest rated player of the team will lead to better matchmaking so I've come around to liking it more as of late. I just kind of wish we had a better solution you know.

Looking forward to hearing your response Phalczen.
DrakolichDrakolích
The BastionTwitch Stream

Verain's Avatar


Verain
02.13.2020 , 12:28 PM | #48
Quote: Originally Posted by phalczen View Post
Should matchmaker take into account "skill," "experience," both, or neither?
I think it should try to find "skill at turning the game into a win". If you can't create a variable that is close to that, then you want to use things like wins over losses, total games played, most mastered ship on current bar, legacy, that sort of thing- basically what it does now, plus W/L.

Quote:
There was a brief discussion somewhere recently about the merits and shortcomings of win percentage as a general, albeit imperfect, measure of "skill."
The merit of this stat is that it is mostly non-cheesable (you can't cheese it in the Wins direction, and cheesing it in the Losses direction results in you losing- you'd need to watch for people who would queue up and afk or float-bot their way to a loss, much as games like League of Legends already do, but you'd probably not need nearly the effort put into it that they have), so selecting for that will be pretty effective. It's not that this is the best stat conceivable, it's merely the best stat we have now.m Obviously, a character with less than 200 games isn't going to have a good way to extract predictive value from this stat, but plenty of players are at far more than 600 games, and their win/loss will definitely have some relation to how well they play. It's not perfect, of course- if you play 10 solo games right now, your own personal skill will likely affect the outcome of one to three of them. Some wins would be a win without you, some losses couldn't have been turned even if you were a walking god. But it's definitely a good measure.

It also becomes a great measure the more someone groups, as it's the only metric we have right now that is VERY noticeable for groups. A solo queue player with a thousand games and a 70% win percent seems to be signalling his skill to the game engine as hard as he can, but the game engine doesn't listen- the distance between 40% and 70% seems to be the greatest information that a solo queuer provides to matchmaker (that matchmaker currently ignores). By contrast, group queues can achieve rates from 85 to 98%, and surely matchmaker could figure out more stuff from this than it does.

Quote:
Accuracy scores could also be a general measure of "skill," and is far less dependent on team dynamics.
This stat is a lot more cheesable. First, cheesing accuracy in the "accuracy" direction is pretty simple. If someone is at the edge of your arc, don't shoot at them. The reason this is annoying is because it's usually correct to throw 10 shots at someone with rapidfire lasers or whatever, because the shots cost you nothing, and they hurt the enemy, but if you are looking at accuracy rates, this is punishing correct play. Second it's easy to cheese with other things- a gunship will have a much higher accuracy than a scout, and anyone focused on only taking shots they don't miss will have a uselessly stat full of worthless puffery. It's also trivial to cheese in the opposite direction- right now you can just exhaust your battery on the way to a node and still have plenty of juice when you need it, and you can really get your accuracy stat low.

The issue here is that whenever you use a useful value as a metric, it becomes no longer a good metric, unless it's something that can't be cheesed.

Examples abound: if I tied your raise to how many type-C forms you fill out, because I know that every customer request yields a type-C form, and I notice that the harder workers help customers more, and I want to increase the amount of help customers get and reward workers who help customers, I'll be flooded with superflous type-C forms. I'll see things that would have been on one form suddenly spread out to four forms. I may even see fradulent type-C forms. I can crack down, but all I'm doing is selecting for the better Type-C Form Spammer/Scammers. The real issue is that my data- the number of type-C forms- was only useful as long as the observed population didn't know I considered it a metric. Since that's inevitable over a long enough time, it really means that I couldn't have used it as a metric.

With the game on live right now, I can guarantee every game ends with an accuracy of above 90 or below 10. Just tell me what I'm going for, you know?

Quote:
Can (kill+assist)-to-death ratios, again imperfect, also be used
Maybe but not really. First, the reward for being high rated has to be assumed to be a universal good, because if player P has a motivation to tank his rating, you just gave him a motivation to self destruct, and he sure can do that. So if you are using it for matchmaking, it's a total non-starter, because some people will want to have a poor rating (they will SD, costing them nothing in their heads, in order to get what they really want, which is matches that they can win), and suddenly they have a trivial way to accomplish that.
Secondly, even if everyone was motivated to make it as high as possible, you suddenly have cases where one team tries to make it so that everyone gets a tag on the enemy, and another team does no such thing. If a team wanted to cheese (K+A)/D, they could do so by maximizing A. You can envision a metric behind the scenes that isn't so trivial to manipulate, but it would be a lot of work, and if discovered, it would likely be cheesable too.

Quote:
Can DOM medals be used
Players don't focus on medals because there's not a huge reward for them, but look at how strange a player who happens to be chasing medals for an achievement behaves. "Give them a node, I need to farm demolisher" (so they need a node to spawn turrets), followed by "let me kill the turret plx I need" and "don't be near node want turret to spawn". Or the player farming healing medals by sc****** their hull down to 80%, waiting for it to heal off their drone, and repeating.
It would also be reasonably simple to scam medals in a downward direction.

Quote:
If any or all of the above four examples are used as indices of "skill," should they be weighted equally, or selectively? Should they be normalized or corrected for the relative balance of DOM v TDM across a pilot's legacy?
If you really just have the existing stats, you'd want a large emphasis on W/L, and a small emphasis on everything else. By making it clear that accuracy has only minor weight, players are less likely to try to scam it, etc. If you put pressure in every direction, a player might have to play really terrible to tank their match-making number, or play really well to increase it, and that's a functioning system.

But you wouldn't want the existing stats. When people look to do ELO rankings in chess or whatever, or ranked warzones in this game, they don't use "medals gained in warzones" or "number of pawns promoted", they use wins and losses, and develop numbers based off of that. Once you have a ranking number, you can do stuff like "this guy is about 1200 goods, and he's up against someone who is 1750 goods, I bet the second guy will win, and we can rig the rewards so that the 1200 guy stands to lose little and gain much, and the 1750 guy is the other way around". But no one is going to figure out superflous stats like "average number of turns to checkmate" or "number of piece-points given up" and build anything on that.

Quote:
I think using "requisition spent on ships on your bar" serves as a somewhat reasonable correction factor to account for a veteran pilot playing on brand spanking new stock starfighters. Should this be weighed equally with the skill factor?
Different. First, requisition needs to count ONE ship- the highest one. It can't average, it can't total. Take the highest ship, ONLY that one counts. That's what it does now, and that's fine.
You can matchmake on this a little, because a player in half mastered best ship will be a little bit less effective than one with a mastered top ship. It isn't a lot, but it's something you can predict and it won't be fully in the player's control. It's ok to have this one.... a little bit.

Quote:
And if this ends up being translated into some numerical score, should each member of a premade be counted individually
Not really. The current system does this because it needs some first order approximation to model that grouping is better than not. If you actually had a real matchmaking system, you would probably also come out with an average in a group that is angled towards the highest member. The current logic (the best guy counts as four copies of himself) would be greatly feared if matchmaker were highly competent and aggressive right now- if a highly ranked pilot joins your group, suddenly you wouldn't be able to get a game you could win, with that logic. Some games have spent some time screwing this up, where they have a functioning matchmaker and they are too punishing towards groups of mixed skill, leaving good players only ever willing to group with each other, or face an unwinnable match.

But as it is, with matchmaker not having the power to do that, the current system is fine. But you would definitely change it if you were doing matchmaker super correctly, it's just a fast approximation that is better than not having anything right now.
"The most despicable person on the GSF forum."

depeshmood's Avatar


depeshmood
02.13.2020 , 03:23 PM | #49
It's nice to see that my idea to open up discussions for how to improve the balance for GSF is keeping traction. I have seen a few suggestions and/or ideas on how to overall improve matchmaking, which is great.
I do hope that the devs could revisit and better refine how matchmaking works. It is something that can actually be felt for both warzones and GSF, but I am really only focused on getting improvements for GSF.

I still think that it would be great if there were more game modes, like the ground warzone 4v4, for GSF.

If only there was a way to know whether or not devs, @EricMusco or @DanielSteed for example, could chime in as to whether or not this would be something they'd even consider revisiting or if we've just been aggravating each other because I asked another player to start this discussion.

Drakkolich's Avatar


Drakkolich
02.13.2020 , 04:23 PM | #50
Quote: Originally Posted by depeshmood View Post
It's nice to see that my idea to open up discussions for how to improve the balance for GSF is keeping traction. I have seen a few suggestions and/or ideas on how to overall improve matchmaking, which is great.
I do hope that the devs could revisit and better refine how matchmaking works. It is something that can actually be felt for both warzones and GSF, but I am really only focused on getting improvements for GSF.

I still think that it would be great if there were more game modes, like the ground warzone 4v4, for GSF.

If only there was a way to know whether or not devs, @EricMusco or @DanielSteed for example, could chime in as to whether or not this would be something they'd even consider revisiting or if we've just been aggravating each other because I asked another player to start this discussion.
Honestly as long as it's a serious back and forth discussion I just enjoy talking about GSF. It's always good to hear different opinions and sometimes it leads to new ideas or strategies. (not strategies in this case since we're talking about matchmaking)

I mean a 4v4 mode for GSF could be cool, I just don't think it should be on the current maps, we'd need like a seperate game mode for that in my opinion. Having played a bunch of 4v4 Death match in Custom games, the strategies can get pretty degenerate currently and the maps are often just a little too big for 4v4's. Now when doing it for fun in Customs it's still hella fun and would encourage many players to give it a shot because you learn so much from those small scale battles, teamwork becomes even more important somehow.
DrakolichDrakolích
The BastionTwitch Stream