Guilty Gear Strive: Open Beta Test Postmortem
An olfactory preview as the game nears its April release
The Guilty Gear Strive Open Beta Test went live at varying times the past week of February 15th, 2021. Select FGC content creators received early beta codes that allowed them to begin playing on Tuesday the 16th; players who preordered the game digitally on PSN got access on Wednesday; and the beta was available for everyone else on PSN on Thursday. After some very unfortunate and frustrating downtime that same Thursday, the beta period was extended until Tuesday, February 23rd at 7am PST.
Some prerequisites before diving into this postmortem: I am by no means a ‘top’ Guilty Gear player. However, I have been playing Gear in some way since 2018, and I’ve also been playing fighting games at a mid-to-high competitive level for most of my life, especially in Pokkén Tournament DX and, to a lesser extent, Super Smash Bros. I’m not going to waste time espousing my credentials again, but you can read my body of work and gain some insight. I also wrote a postmortem about the April 2020 GGST CBT, but my knowledge both about Gear as a whole and GGST has changed drastically since then. In the (relative to other players) short time I’ve played Gear, I’ve grown attached to the game’s wild heavy metal anime aesthetic and near-infinite level of player expression, such that I would say it’s now my favorite fighting game IP. I’ve put several hundred hours into learning how to do things with my character that used to hurt my hands when I practiced them during the first dozen. Guilty Gear is sick as hell.
With that out of the way, let’s dive into Strive. I will say that my early impressions of GGST are positive, probably more so than my fellow Gear heads, but I still have some issues and criticisms. I highly doubt that any major game engine changes will be made in time by the April 9th launch in a month or so, but I’m sure that after several characters and balance revisions later down the line, Team Red will tune the game more to my (and other players’) liking. Furthermore, I’m honestly fine with playing the day 1 product for a long period, assuming it essentially resembles the version I played this past week. That said, certain elements of the game definitely require some immediate addressing and major quality of life improvements, which I will obviously be pointing out in this article.
Probably half of my time spent with the game was shared on my Twitch stream, with the other half being me experimenting and grinding the Tower system privately or playing with friends. I mostly played Giovanna, with a little bit of Sol thrown in, and I ended my beta Tower climb at floor 9. In my CBT article, I broke down my observations from day-to-day, but since all of the features of the game were available from the first day and the game itself was straight-up unplayable for a cumulative 24 hours, I’m not going to do that this time. Instead, I’ll be going through the feature list (to summarize for people who may have missed some things or didn’t get to play the beta at all) and then I’ll break down things I’m opinionated on.
Phew! Lengthy preamble! I just want to make sure everyone knows what they’re getting into.
GGST: What’s in the box?
The beta had a handful of features that a lot of people probably didn’t try out because they were busy either in Training Mode or playing online. Singleplayer and Story Mode content were not ready in time for the beta, but you could still play against CPUs or against local opponents in Vs. Mode. Vs. Mode obviously has all the features you would expect from an offline head-to-head 1v1 mode, including an instant rematch button and a UI element at the Results Screen that keeps track of wins and losses over the length of a long set. (This is important for later.) There was also a Tutorial Mode that served as a framing device for the game’s story, as Ky trains Sol properly for the latter’s inevitable fight against That Man. (If you have no idea what I’m talking about, ASW has all the meat of Gear’s story on their YouTube and on the GGST official site via a short e-manga.) The Training Mode was absurdly robust, with almost every feature you could want out of it including dummy recording for reactions and reversals. The only feature I didn’t see in the mode, which was disappointing, was a frame data and/or hitbox viewer, since even SFV has that now and GBVS at least let you see vaguely whether you were plus, even, or minus after a move. I assume that this would be an easy addition and that it may not have been included in this build since combat data is subject to change prior to day 1, but obviously I would prefer not to be checking Dustloop for frame data constantly while labbing.
Much like in the last beta, the first thing you do when hopping online is create your pixel avatar and get matched up with a CPU who then assigns you a rank placement. (Fun fact! My original CBT pixel avatar has been used in Sajam and Jiyuna videos without them knowing they got it from me. It’s a long story.) I was initially placed in the fourth floor before eventually climbing to the ninth. Newly introduced to the beta were three major features: the Quick Match button, the Park lobbies, and the ability to enter Training Mode while waiting for a match. The Quick Match button lets you enter a ranked Tower floor without actually having to step into the pixel lobby properly: by placing your avatar in the lobby automatically, it allows you to queue up against players who are wandering the lobby as well as any others who are Quick Matching. The Park lobbies are an unranked alternative to the Tower floors, notable in that each matchmaking region has a set amount of lobbies with a 32-player capacity. If you wanted to play with your friends during the OBT, this is usually where you went, since the actual private player room system had not yet been implemented. Finally, the Training Mode standby was functionally similar to the Quick Match button, except it didn’t put you into the ranked queue when you selected it while in a Park. All of this was rounded out by the widely-anticipated inclusion of ASW’s newly proprietary rollback netcode implementation, which Akira Katano even mildly boasted about at the second Japan Fighting Game Publisher Roundtable this past Saturday.
There were a handful of other miscellaneous features as well. If you were ever curious about the standings of yourself or other players during the OBT, you could check Rankings and see for yourself. You could also view replays- again, both yours and other players’- in the Replay Theater, where all replays are uploaded to a server. (Online matches have their replays saved by default, which can be turned off.) Finally, there is of course a handful of system options, including things like audio and button config.
The nitty gritty
Let’s just get this out of the way: The rollback netcode is fucking phenomenal. The rollback alone straight-up makes me never want to touch Xrd ever again. I’m tired of fighting games with online play that don’t respect the time I put into learning the character or the systems in training or in offline play. I just don’t have that patience anymore. GGST’s rollback, built from the ground up for this game, and allegedly taking queues from GGPO as a framework (again, according to Katano of ASW) makes online play feel so much more satisfying than nearly any other game I’ve played. Between GGST, +R, and maybe one or two other games like Them’s Fightin’ Herds and Skullgirls, I don’t want to ever play another fighting game with delay-based netcode. KOFXV later this year is on notice. While server-hopping in GGST, I was able to play games with European players at 180ms or higher with no input delay and very minimal visual or audio artifacting. I even played one of my friends from Australia at 225ms and it was so incredibly smooth. Rollback might not be a magic spell, but it certainly damn feels like it.
Now let’s move onto the combat. I mainly played Giovanna in this beta for two reasons. One, because superficially she seemed like the type of rushdown brawler archetype that Jam fulfilled for me in prior games (although in reality she’s clearly more like Cammy than she is Chun-Li.) Two, I figured that trying to learn the GGST engine through an entirely new character would help me parse my muscle memory better. I didn’t try Nagoriyuki for that long, and I also did experiment a bit with Sol, who was the character I played exclusively in the CBT. Giovanna herself feels super fun. Rushing down the opponent with her very strong S buttons out of her step dash felt natural and strong. Said step dash, along with her Trovao special move (the projectile-invulnerable Spiral Arrow move) let her move forward across the screen very quickly and I rarely felt I had a tough time closing in on opponents outside of a handful of matchups. I never felt like I had to conserve my meter with Giovanna despite her self-buff mechanic, since her Ventania super was useful as a reversal and the Roman Cancel system in GGST was intricate and powerful. While I think she’s still missing a handful of options- more air OK special moves besides Sol Poente, and a better invulnerable anti-air besides 6P- she is a generally satisfying rushdown character.
Playing as Giovanna helped me parse a bit more about what I liked and didn’t like about the GGST gameplay, and while I’ve seen a lot of valuable opinions elsewhere I don’t think it’s as drastically different as other people may think. GGST is still undeniably Guilty Gear- it’s just that the possibility space has been narrowed. Some options aren’t as weak as people initially assumed, however. Jumping in and airdashing forward, for instance, are still pretty strong options and make for an interesting air-based game. Likewise, I wouldn’t necessarily argue that anti-airing has been reduced in power, either, although some characters like Giovanna and even Sol have weak or at the very least inconsistent anti-air options and follow-ups. The main issue that I’ve found instead is that air-blocking without Faultless Defense has just made jumping backward really good, especially when escaping pressure or the corner. It’s not always easy to chase an up-backing opponent, and air-throw becomes a very telegraphed option in this situation. Having said all of that, GGST is still a typical anime airdasher with ‘air footsies’ and intricate combat on the top half of the screen. I also don’t mind things like auto-tech/no air tech replacing mash-tech, although I would have probably preferred a hold-tech option.
Incidentally, my theory as for why airdashes are the way they are in this game is based on having read the Infil article on rollback. In said article, Keits suggests that IADs should have a few more frames of additional startup attached to them while preserving the overall inertia of the airdash. (He cites Marvel vs. Capcom Infinite as his example, pointing out that Spider-Man and Chun-Li experience a lot of rollback when moving through the air in high ping.) If you’ve ever played GGPO+R, you’ve probably seen characters like Chipp, Bridget, and Jam warp around in the air when playing a connection with high ping. Even though +R’s rollback is incredibly designed, the game wasn’t built with rollback in mind, so certain visual artifacts like that happen frequently. Compare +R to GGST, a game with a lot of front-loaded startup animation or moves with a lot of hitstop- there will be very few instances with any visual artifacting. It probably helps that CMZinac, a former Killer Instinct developer and colleague of Keits, was working on GGST’s netcode and likely influenced the combat design in this direction.
My main issues with the game are essentially the same as they were in the CBT:
- Gatlings being limited in their application doesn’t feel too great,
- the wallbreak doesn’t feel like much of a ‘choice’ for most characters,
- and the damage is way too high.
For the unfamiliar: Gatling combos are essentially Guilty Gear and Blazblue’s version of the Vampire and vs. Series’ magic system, or Reverse Beat in French Bread games like Melty Blood or Under Night. The premise is simple. You have five attack buttons: Punch, Kick, Slash, Heavy Slash, and Dust. Of those five buttons, you can cancel P, K, S, and H into each other in nearly any ascending order for an easy-bake combo. (You can do all four, PKSH, you can start later, KSH, or you can even skip a tier, PSH.) Chaining combos like this has been fairly intuitive in previous games, and has also historically been a very beginner-friendly system. Anybody can get a combo this way, just by rolling their hand along their controller in a specific order.
In GGST, the Gatlings have been reduced substantially. P and K no longer chain into each other. Instead, P chains into itself, command normals, and special moves. K does not always chain into itself, but can be jump cancelled or chained into D. S and H function as usual. This does open a handful of new options- mainly, I’ve been using 5K and 2K as set-ups into sweeps, tick-throws, or even Dust launchers. In general, though, this greatly limits the utility of your light buttons, both as starters and in the middle of combos. This generally makes combos shorter, and your combo starters are now harder to earn or higher committal. If you don’t have a blessed command normal or an especially strong c.S, you aren’t going to get much of anything in neutral, and it will be difficult to extend your combos without meter. Overall, your character expression feels more limited and restrained in GGST than it did in prior Gears because of this change. Even though some characters have had their kits reduced- Chipp and May in particular- the universal Gatling change is more significant in contributing to this feeling of restraint.
Beyond that, there’s also the matter of both the wallbreak being nearly unavoidable and damage being way too high, both of which go hand in hand. While okizeme and setplay are still present in the game (just look at Millia, Zato, and even Ram’s reworked kit) the wallbreak mechanic does a lot to reduce those options across the board. On any hit that wallsplats the opponent just before the wallbreak, you have two choices: let the opponent drop from the wall, or break the wall. Compare the disparities in what you earn with each choice. If you let your opponent drop, they don’t slide down the wall like they sometimes did in past Gears- instead, they tech off the wall and neutral is reset. If you wallbreak, on the other hand, neutral is still reset, but you get a bit of extra damage, the Positive Bonus damage and meter buff, and a knockdown if you broke the wall with a super. On top of all that, the wall health doesn’t reset if you don’t break it, so there’s very little decision making to be made here. In the current version of GGST, there’s no incentive not to break the wall. You even get more okizeme potential for breaking the wall with a super than you do for letting the opponent drop from the wall!
This probably wouldn’t even be a big deal if damage wasn’t so inordinately high. Guts is still present in the game (although it’s hard to tell, but Guts seems to be in effect once your health bar turns orange) but seems to be really toned down, so there’s not a lot of damage reduction later in the round. Furthermore, any eight-to-ten-hit combo can deal roughly 200 damage, meaning your opponent is half dead or worse before the wall has even been broken. The reward you get for breaking the wall seems almost meaningless given the fact that you worked so hard to nearly KO your opponent just to get them to the corner in the first place, and since wallbreak doesn’t always kill outright it also means you have to break the wall and engage in a last-hit scramble instead of maybe getting some meaningful okizeme like you would in past games.
In spite of all of this, though, the game itself was still fun to play and felt like I was playing Guilty Gear. I still had long combos and pressure and was able to express myself at every part of the screen, especially once I got the hang of the new Roman Cancel systems. Blue RC feels like a much more tuned version of YRC, and the ‘Drift’ and ‘Fast’ RCs let me do a lot of creative mixups and combo extensions once I got the hang of it. (Using the dash macro helped a ton.) I also wasn’t especially hurt by the changes to okizeme and setplay, but I will acknowledge that my character choice probably contributed to that a lot. Overall, I think that the engine mechanics have the potential to be some of the best in the Guilty Gear series, especially with the new inventive RC system, but it would take restorations to the Gatling system at a minimum to get there. Wallbreak itself can remain in the game- it’s visually and technically impressive, it’s got the wow factor for spectators, and there may be situations down the line where you want Positive Bonus- but I think for it to be a unique choice, the wall interactions need to be changed in a way that facilitates resets into okizeme. Damage being tuned down is fairly straightforward, but could also happen as a consequence of either of the prior mechanics being tweaked- perhaps restored Gatlings means more damage scaling and combo proration, or the damage prior to a wallbreak is significantly reduced. Either way, I like playing GGST despite my personal hangups, which is good.
The lobby
A universal complaint across the board has been the pixel lobbies- I take exception only to some of the complaints. I actually rather enjoy the presentation of the lobbies. I like being able to jump around with my dressed-up little guy and hang out in a room, swinging my metal bat at people. Moreover, I think the rank system works really well: it prevents newbies from being deliberately stomped all the time, but newbies themselves can challenge the upper crust of the game if they so choose. Rewarding players for hitting the hidden eleventh floor with a prestige rank is also a nice incentive. The main problem with the lobbies is just that their feature set doesn’t seem to work that well half the time. (I’m going to exclude issues that are obvious bugs from my complaints, such as desyncs during matchmaking or being kicked out of Training Mode after a failed match initialization. While these are justifiable complaints, they’re not inherently brought about by a particular feature set, which is what I want to focus on here.)
A major part of the pixel lobby that had been changed seemed to be in direct response to how broken they were during the CBT. Now, instead of readying your avatar’s weapon and touching someone who was also in standby mode, you enter standby with a held button press and are forced to remain in place. This would normally not be an issue if it weren’t for the fact that, upon entering standby, your avatar warps away from where you were initially standing, often-times onto the other side of the lobby. The warping is also probably ASW’s response to the issue of overlapping avatars in the previous build, but unfortunately this still happens in the OBT. These issues can make your character difficult for other opponents to find, even when making use of the Member List option to teleport directly to you. Beyond that, it’s not always easy to tell when someone has already begun a match with someone else, meaning you can often queue up with someone and then get kicked out because someone got there first. (It’s easier to see this when there are avatars that are Quick Matching. A number will appear above each avatars’ heads, counting down from 15 in sync, before the two of them eventually pair up.)
The easiest way to solve this issue would be to simply bring back the arcade cabinet kiosks from previous games. They can retain the pixel lobby look if they really want to- again, I really don’t think the presentation is as big a deal as everyone says. But the kiosks are an unambiguous method of communication. They’re a fixed location and their message is clear- just go to one of the cabinets and sit down if you’re looking for a match. It’s also easy to tell if someone is looking for a match- the seat next to them will be empty. Trying to ready up, being warped around the lobby, and failing to connect to someone because someone else got there first in the time it took for you to find them inevitably means you spend longer in menus than in an actual game. Quick Matching mitigates this somewhat, since, contrary to popular belief, it does in fact pair you with other Quick Match opponents as well as offer you up to players wandering the lobby. (It just so happens that the aforementioned desync bugs made Quick Match in the beta somewhat annoying.) Still, the absence of the kiosks given their presence in Xrd and GBVS is really bizarre. ASW can even let avatars keep the weapons they swing around- just let players continue to cosmetically swing them with the Slash button instead of using them as a standby indicator. Adding the kiosks in rather than overhauling the entire lobby system is also probably an easier change.
Beyond that, the fact that there was no instant rematch button for the beta sucked for the same reasons as above. No instant rematch in public lobbies, either in the Tower or in the Park, meant that you had to do the standby song-and-dance again, load back into the game and sit through the loading eagle (an even bigger pain if you’re not playing on PS5, since load times on PS4 and PS4 Pro were pretty long) and then fight. I can somewhat forgive this feature missing from the beta for a few reasons. One, the button itself is already in the game- see that I mentioned it above when discussing Vs. Mode. It would just require a toggle for the final version of the game for online modes. Two, I assume that the instant rematch button is just one of many features excluded from the beta in order to encourage players to seek out a variety of matches with many people. (While ASW hasn’t specified what exactly these features were, they did mention during the JFGPR that a Wi-Fi indicator would be one such feature.) Three, there will almost assuredly be instant rematch button or an option for longer sets in the private player rooms that will be included in the game at launch. (If I’m wrong about any of this I will eat a shoe or something, and also frankly be very upset.)
The online lobby system is probably a bigger barrier to getting into the game than any gameplay mechanic. While game feel may be compromised to veteran players, you can ultimately adjust to it over time and find value in it as you play more and experiment. But you can’t overcome that game feel being ‘off’ if you don’t get to play the game in the first place. GGST might be fun, and the netcode might be perfect, but it’s locked behind a dysfunctional lobby that needs a ton of tweaks ahead of the April launch.
Conclusion
Despite the issues I outlined with the game here, I played the GGST beta for probably a few dozen hours and I had a ton of fun. I’m still looking forward to the game and overall its online is a major improvement over Xrd. Any game that makes the online and offline experience feel identical is going to reward players for their time investment to a substantial degree, and will thus be a more valuable competitive experience as we move forward into another year without offline FGC events. When I filled out my GGST OBT survey, the main issues I focused on were those I considered a priority: the restoration of Gatlings, the improvement of the online lobbies, matchmaking filters and instant rematch, and a future implementation of PSN-Steam crossplay.
GGST is a wildly experimental game, even by Arc System Works standards. It might not be everyone’s thing at launch, but it’s also succeeded at getting a lot of new people interested in Guilty Gear. People in my Twitch community and close friends who were turned off by Gear prior to Strive are suddenly excited to play the game again in April. Maybe certain systems in previous games were difficult for them to understand; maybe the new battle changes allow them to pick up and play the game easier. While the possibility space of previous Gears is incredibly vast, not every player can overcome the barriers that prevent them from accessing that bottomless well- but perhaps they still want a game that challenges them just enough. Either way, they’ve all been telling me how hype they are to play GGST, and they’ll be able to do so without having to choose from a very limited pool of opponents with ping lower than 100ms. I already know I’ll get a lot of enjoyment and personal growth out of the product on day one. For everyone else, I have faith that, much like previous Gears, the game will become something closer-to-perfect over time as the community gets used to it and the developers tune their experiment in the right direction.
Also, frankly, if you think the vocal OST sucks, you can eat my entire ass.
