None of these titles made a significant mark on the competitive scene. None of them really mattered. Yet somehow the fighting game community survived for almost a decade without a new release worth supporting in tournament play.

We steadfastly, stubbornly stuck by MvC2, SF3:3S, CvS2, and SSF2T – somehow managing to help grow Evo attendance numbers each and every year. That’s because one of the most remarkable things about the fighting game community is that when a new version of Street Fighter is bad, we know to play the old one instead. Pretty simple, right?

In retrospect, this approach was nothing short of miraculous. SFA2 and SFA3 got us through the disappointment of the SF3 series. MvC2 kept the community strong through the debacle that was CvS1, until CvS2 repaired most of the damage done by CvS1. Then SF3:3S started becoming good and carrying its own weight. And Super Turbo was there all along.

Then we were given nothing for eight years. It would’ve been so easy to become disillusioned and quit, but enough of us stuck together to define our hobby on our own terms. Let’s not lose that culture of self-reliance now. We don’t have to keep playing Street Fighter X Tekken or any new game if it doesn’t meet our standards.

Don’t get me wrong – SFxT is still a great product for fans of the SF4 series. It’s all your favorite characters thrown into an entirely new setting, paired with awesome rivals from another franchise. That makes SFxT fun to explore in the same vein as CvS2 players enjoyed the chance to try Aegis Reflector combos with Guile in Capcom Fighting Evolution.

Unfortunately the reality is, it takes an extreme level of polish (and a huge dose of luck) to put out a solid tournament-quality fighting game. Personally i think it’s worth supporting Capcom for taking an honest stab at it. You should totally buy SFxT and try it out for a month or two!

However if the game simply can’t hold up to rigorous competitive play, there’s nothing wrong with shifting our attention back to the last version that passed the test. The SF4 series might have gone a little overboard with comeback mechanics and option select madness, but otherwise Super Street Fighter IV Arcade Edition Ver. 2012 is a truly legit tournament game.

So if you’re tired of SFxT’s gems and infinites, why not go back to SF4? Or try Skullgirls. Just remember, the choice is ours.

PlayStation All-Stars Battle Royale is a new four-player fighting game developed by SuperBot Entertainment, studio presidented by Chan Park and game directored by Omar Kendall. I’ve been working in the combat design department under lead combat designer Paul Edwards.

We’ve grown into a fairly large team – with quite a few oldschool fighting game community members whose names you may recognize, such as senior combat designer Ed Ma and lead animator Tommy Ho. These guys have been in the game industry for a very long time and their experience has been a huge asset on this project.

Our goal was to make every facet of Battle Royale’s gameplay as deep, varied, polished, and accessible as possible – so that it would be fun and rewarding for anyone to pick up and play. When it comes to core combat systems and high level design decisions, i have a ridiculous amount of faith in Omar, Paul, Ed, and our team as a whole.

For more information about the first six playable characters and first four levels, please check out any of the following hands-on previews:

GT.TV | IGN | G4 | GameSpot | Kotaku | Joystiq | TotalPlayStation | USA Today

I really can’t wait to see where the fighting game community takes PSASBR when you get your hands on it this holiday season. I hope you all thoroughly enjoy playing it.

0:11 The entire goal of this setup was to integrate two F+PPP Yoga Teleports into a single combo. Seth’s teleport lasts for 40 frames while his s.HP only causes 22 frames of hit stun. How do you fill the gap? Put a Sonic Boom in the middle! Unfortunately, Seth’s LP Sonic Boom travels annoyingly fast whereas his walk/dash/jump movement speed sucks. Furthermore, the SF4 camera resists scrolling which drastically impedes Seth’s ability to race past his own fireball. Luckily it turns out that M.Bison’s HK Knee Press Nightmare pushes (slightly) past the edge of the screen, which creates that tiny bit of extra space necessary to delay the fireball impact enough to make this sequence click. Of course, the SF4 engine hates crossup combos so there isn’t much Seth can do midscreen with Bison leaning into him. I thought s.MK would look more interesting than another s.LP, c.MK rerun. By the way, that c.HK sweep at the end is a one-frame link off lvl2 Focus Attack xx dash forward, F+PPP Yoga Teleport.

0:22 Trading with Seth’s Hyakuretsukyaku produces roughly the same hit stun duration as a standard hard attack, but interrupting him shortly after it connects maintains its unique extended reaction. Both instances are shown here. First, Seth’s LK Hyakuretsukyaku trades with Rose’s close s.LK, then her LP Soul Spark interrupts another LK Hyakuretsukyaku to set up crossup j.HK, c.HP, HK Hyakuretsukyaku back into the corner for the stompy super finish.

0:40 “Maj Productions presents Seven Focus Attacks! (starring Seth’s Focus Attack as Chariot Tackle and Yoga Catastrophe as Aegis Reflector).” I’d been meaning to revisit this setup with Seth ever since i introduced it in SF4 TACV 02 using El Fuerte, but error1 beat me to it. Nevertheless, i decided to include my own variation with seven lvl2 Focus Attacks, mainly for the sake of completeness. I kept the combo out of the corner as long as possible, but it’s necessary for the fifth Focus Attack to connect after the first HP Sonic Boom. Also, Seth can perform lvl2 Focus Attack xx dash forward, lvl1 Focus Attack as a ground combo, so theoretically it should be possible to KO with that lvl1 Focus Attack and juggle another lvl1 Focus Attack. Sadly, it’s a couple frames too slow to connect. This combo inflicts 318 damage and 534 stun.

0:59 As if Seth’s dizzy potential wasn’t already through the roof, Pokey86 discovered he can link close s.HP after close c.HK under certain circumstances against a handful of characters. It works after j.HP on crouching Sakura, who happens to have 1000 stun points – not a problem for someone who can begin a combo with five hard attacks in six hits. The post-dizzy combo simply showcases Seth’s ability to juggle j.D+MK after c.HP xx MP Shoryuken x2, which requires surprisingly precise spacing against Sakura. This combo deals 571 damage.

1:21 This combo aims to maximize Seth’s damage output by leveraging his newfound ability to link close s.HK, close s.HP after Tanden Engine on standing Seth or j.HK on crouching Seth, which leads to 785 stun in four moves without meter cost. Once he reaches the 10% damage scaling threshold, close s.HK, close s.LK, s.LP, c.MK xx HP Sonic Boom xx lvl2 EX Focus Attack becomes his most damaging ground sequence. Apart from containing several one-frame links, this combo is relatively straightward, yielding 642 damage.

1:46 All versions of Seth’s Tanden Engine elicit the same reaction on success without inflicting damage or registering on the combo counter. Seth himself recovers when the opponent touches him. The result is that Seth gains +23 frame advantage on hit at point blank range, and progressively less frame advantage the further the opponent must travel to reach him. Seth’s EX Tanden Engine acts as a simple hit, which means it’s capable of affecting opponents during hit stun. Conversely, the regular versions of Tanden Engine behave more like throws, so they whiff through hit stun. EX Tanden Engine also happens to become active on frame 23, thereby making LP Tanden Engine, EX Tanden Engine, EX Tanden Engine a genuine (zero-hit, zero-damage) combo with two one-frame links. Killing an opponent with a light attack renders them vulnerable to throws which also includes HP Tanden Engine. The only reason Honda serves as the opponent is because Seth can whiff close s.LP twice over his head after EX Tanden Engine and still have enough time to link c.LP at point blank range, bringing the total number of unregistered attacks up to six.

2:06 Seth’s lvl3 Focus Attack trades with Seth’s lvl1 Focus Attack because it looks cool, which combos into lvl2 Focus Attack, which sets up a full-arc j.D+MK pseudo-OTG which looks cool. Landing the crossup EX Tanden Engine requires a particular combination of j.D+MK stomp timings, but it’s fairly straightforward otherwise. Seth’s body bobbles slightly during his focus crumple animation, which acts as a meaty setup for s.LP, providing the 1 frame of additional advantage necessary for s.LP, j.D+MK -> j.D+MK -> j.D+HK to link properly. Seth’s height and hit stun width cast him as the perfect dummy candidate for this combo. However, the j.D+HK divekick connects too high, making it impossible to follow up normally. Luckily j.D+HK happens to trigger a (slightly extended) light attack reaction, which means it causes crumple instead of knockdown on KO – allowing both hits of s.HK to pelt dead Seth.

This is a tool-assisted combo video recorded with two ASCII PAD V Pro programmable controllers. Absolutely no cheats, hacks, or game-altering devices were used in the making of.

Skullgirls is a truly hardcore fighting game which pulls no punches when it comes to offering challenging gameplay. Yet beneath the surface, it’s extremely clever about staying accessible and streamlining basics to minimize drudgery.

As you delve into exploring Skullgirls, you’ll notice that the smallest details feel like someone put genuine thought and effort into improving conventional designs. Air dashing, throw teching, input buffering, infinite avoidance, assist selection, and even 360 commands have been brilliantly implemented – with subtle innovations that you’ll be delighted to uncover.

In this day and age of developers dumbing down the genre to conceal that steep learning curve from new players, Reverge Labs has taken a huge risk by making a title for fighting game veterans – and i think the community should support them for it.

If you really want to play the game well, you’re going to have to get your hands dirty and learn all those scary things that make fighting games rewarding. The interactive tutorial will teach you the essentials as quickly as possible, but that won’t keep better players from putting you in frightening situations that only skill and experience can overcome. There’s no escaping that simple fact, but isn’t that what the genre is all about?

I suppose it wasn’t meant to be, but i can’t help feeling a little disappointed – only because the script they showed me was outstanding. I can only hope they get another opportunity to make the film a reality someday. However, there is a silver lining here:

Continue? Episode 9 | Episode 8 | Episode 7

Continue? is a ten-episode web series starring Greg and Brett themselves, with three episodes released on the CBF u2b channel so far. The first one covers the premise and inception of the show, the second episode hits the ground running with some hilarious storytelling, and the third is essentially their reaction to the end of the Matador campaign.

Each installment will count down the clock until it reaches zero, at which point they will reevaluate if it’s worth continuing any further. So if you like the series and want to see it extended beyond ten episodes, please take a minute to show your support by subscribing to their channel or following @ChubbyBoyFilms on twitter.

Obviously the fighting game community has some growing up to do. We ought to accept the challenge because it’s the right thing to do, not because it matters what sponsors or gaming journalists think of us. It’s fun to hear top players talking trash, but we can all tell when lines are being crossed. If you see it happening in front of you, do something about it.

Being an instigator is easy, but you get nothing out of it. Anyone can be a stream monster. Acting as the voice of reason is far more difficult, but infinitely more rewarding.

Will the entertainment district of the internet ever become politically correct? Probably not. Online forums and anonymous chatrooms are always going to stay a little sketchy, so there’s no need to champion the fighting game community’s right to be grimy.

Let’s be honest – the FGC tournament scene will probably never turn squeaky clean, so it’s important to pick your battles. Just don’t forget or ignore where the real lines are drawn and don’t be afraid to speak up if someone goes too far. That’s all anyone’s asking of us.

As a spectator, i feel bad about what happened to SuperYan specifically because she seems like a genuinely nice person who was there to play the actual game of SFxT. It was entirely unfair to make her feel so uncomfortable. That’s the whole reason there aren’t more girls in the fighting game community, which in turn is the reason they get driven out like this.

It’s a pretty messed up vicious cycle, because it’s hard for anyone to cope with that much critical attention suddenly directed toward them. I don’t know how SherryJenix has managed to handle it so well, but it would be completely unreasonable to expect such restraint from everyone who simply wants to participate in our community.

SuperYan, if you still want to get better at Street Fighter, i’m down to help. When you feel ready, record a few videos against friends and post the links here. I’ll take one match per week and give you as much constructive advice as i can. Somebody from Australia once paid me $40 for feedback on two of his matchvids, and he seemed happy with the outcome. I’m pretty sure that makes me a professional Street Fighter tutor.

OHNX CMV

Featuring contributions from eleven well-known combo makers, this video contains a wide assortment of classic and modern fighting games. Usually these projects take several months to come together, but somehow CPS2 got it done in 2-3 weeks – and it turned out great!

The two ST Guile combos were recorded as test clips a long while ago. The one at 1:02 was the prototype for the Guile mirror combo at 7:29 in SF? Guile Exhibition – minus the Time Over setup. In case anyone was wondering how it would look without that setup, now you know. The other clip at 4:02 was the most i could get out of a backwards super vs Zangief.

I also recorded the CvS and CFE combos based on some ideas that Keiko and i brainstormed, but i don’t have much to say about those clips – except that Izuna Dropping dinosaurs is cool.

0:02 Dan makes an exceptional combo dummy in CvS2 because his Chohatsu Densetsu super bestows 1/8 meter with each of its first five Taunts, then 3/8 meter from the final Taunt. Thanks to that hefty refund, C-Morrigan gains access to four bars of meter. Her lvl1 Soul Phoenix connects after she initiates lvl2 Valkyrie Turn, fulfilling its cancelability requirements. Canceling into lvl1 air Soul Phoenix then creates a free juggle state and resets the juggle counter, allowing both hits to connect. Valkyrie Turn also locks the screen in place until Morrigan reaches neutral state, thereby creating a temporary pseudo-corner situation.

0:14 Kyo’s lvl3 Serpent Wave barely catches Iori’s descent from an air-to-air j.D+HP sudden knockdown. OTG LP Flashing Slash connects at maximum distance in the corner.

0:20 Todo’s crossup j.LK triggers Blanka’s Counter Attack facing the wrong direction, then F+MK slides into its last active frame. Todo’s momentum carries him over Blanka’s lvl2 Shout of Earth, which can only be mashed using one kick to avoid unintended special moves. Canceling into whiff LP Rolling Attack creates a free juggle state for vertical j.HP to connect.

0:25 Yamazaki’s Taunt & Counter clips Bison’s feet as HP Psycho Crusher passes through, knocking Bison back over Yamazaki. I found this cool-looking midscreen setup by accident, but unfortunately Yamazaki doesn’t have much in the way of viable follow-ups.

0:30 Normally, Todo’s lvl3 Ultimate Ecstasy Crunch is only supposed to juggle up to nine hits. Geese gains some extra altitude by sticking out j.MK from a distance, and eventually starts falling into the waves two at a time. Simultaneous hits only count once against juggle potential, which explains the three extra hits.

0:35 Blanka’s HK Vertical Rolling trades with Chun Li’s crossup j.LK, which provides barely enough air time to recharge for lvl2 Shout of Earth. The following c.HK juggle is made possible by Chun Li’s special falling hit box, creating just enough charge time for lvl1 Ground Shave Rolling. Two charge direction switches occur during this combo – one before each super.

0:41 Kyo triggers Maki’s Counter Attack glitch to demonstrate the hidden juggle potential of Seven Hurting Combos. Precise corner spacing acts as a meaty setup into lvl3 Serpent Wave.

0:49 Akuma chases after his lvl1 Messatsu Gouhadou while Dan’s air Taunts refund his meter. The fourth Taunt was probably unnecessary, but it’s shown here in case anyone needs it for another combo. Akuma’s Hyakki Goudan knockdown is nullified by connecting simultaneously with his super fireball, allowing LK Tatsumaki Zankukyaku to juggle twice. Dan is susceptible to CC mode OTG throw attempts, including Akuma’s Hyakki Goushou.

1:04 This kid in Germany used to send me e-mails about CvS2 combos all the time, going by the name DerT7. As far as i know, he moved to England and stopped playing fighting games, but some of his ideas were pretty clever. He showed me a basic version of this combo starter quite a few years ago, following up Yun’s lvl3 You Hou with his lvl3 Hiten Souryujin. I took this opportunity to spruce it up with some crossup juggle madness.

1:21 Canceling Akuma’s lvl2 Tenma Gouzanku into Tatsumaki Zankukyaku immediately after superfreeze makes him land abruptly instead of floating upward. There’s just enough time to tack on s.LP xx HP Goushoryuken amid the fireworks. This discovery was made by jchensor, and i’ve been meaning to incorpoate it into a combovid for a long time now.

1:27 My friend omni found that Akuma’s lvl2 Messatsu Goushoryu sometimes slips under opponents as it juggles after Tatsumaki Zankukyaku. Of course this means it can set up a backwards attack under certain circumstances, which even leads to a midscreen dizzy combo.

1:40 Here’s another old combo i’d been wanting to put in a video forever, and it happens to be the only previously unreleased clip from this batch. Squeamish and ALTANertive showed me their awesome Vice midscreen CC loop, way back in 2003. She whiffs far s.HP to reposition herself into HK Da Cide range midscreen. I added the HK Da Cide Slayer trade against Rolento’s LK Stinger for a little extra flavor. I can’t believe it took nine years to get this out!

1:53 Zangief’s elaborate corner-specific multi-cross CC ends with lvl1 Aerial Russian Slam. Somehow he’s able follow up with OTG HK Atomic Suplex – but only for one frame after CC, only against Dan and Rolento, and only if they’re already touching the ground plane. Choreographing the crossup juggles took a lot of work, but they were necessary to place Gief in the corner. I discovered a long time ago that the left corner lets characters pop out during certain juggle setups, so i have j.LP, walk forward, PPP Double Lariat xx s.LP, j.D+HP start the crossup jumping sequence. If i had let Dan go after the Lariat, he’d fall way too far away for poor Gief to catch up. Instead, Double Lariat knocks Dan out of the corner and s.LP pulls him back in, but the problem is that Dan still has corner position priority, which means even if Gief is walking back with absolutely zero room between him and the corner, Dan will still fall back into the corner. That’s why Gief waits until the very last frame before he would turn around and jumps forward with crossup j.D+HP to pull Dan out of the corner again. From there it’s more or less clear what happens – except for the OTG command throw, which is just crazy.

This is a tool-assisted combo video recorded with two ASCII PAD V Pro programmable controllers. Absolutely no cheats, hacks, or game-altering devices were used in the making of.

Fancy words aside, the main problem for me is the amount of braindead grinding needed to access the coolest toys in the game. Naturally we’re all going to find some snazzy piece of armor and want to complete the rest – or hear about a rare pet/hireling that we can’t resist. Should it cost fifty hours of repetitive chores to unlock these scraps?

Action RPGs tend to be engaging at the beginning because we’re constantly seeing new things and making meaningful decisions at a fast pace. Ultimately that rhythm grinds to a halt, and the amount of dead time between exciting new discoveries stretches into hours. Daring to want that last component of Griswold’s Legacy means killing a thousand recycled monsters beyond the last unfamiliar pixel you’ll ever see.

After i finished all of the story quests in Torchlight and reached lvl35 with my first character, i discovered that the fastest way to obtain unique items was by scamming the gambler NPC in town. I promptly maxed out my Barter skill, equipped as much Decrease in Vendor Prices gear as i could round up, and filled my inventory with stacks of Dogfish/Catfish. Then i sold, bought, and resold the entire stash hundreds of times, twenty clicks at a time, and used the profits to buy irregularly priced socketed unidentified items from Duros the Blade at a steep discount. Over 1/3 of them were uniques, yet it still took forever and a half to assemble Magnus’ Trove.

It was probably the least fun i’ve ever had playing a video game. And then there’s the fishing. Don’t get me started about the fishing. How can any of this fluff be considered gameplay? Don’t these arbitrarily time-consuming activites seem like they were contrived purely for the sake of boasting X number of replay hours? Personally i find it disingenuous, and i think it tarnishes an otherwise excellent game. There has to be a better way, right? I certainly don’t have all the answers, but i can offer a few suggestions.

What if completing 150% of the game’s non-repetitive play experience enabled direct access to all hidden prizes in the game? For instance, if the average player finishes the main story with all unique side quests at character level 35, then someone who reaches lvl50 is probably searching for something specific, right? Why not let them have what they want?

It wouldn’t have to be free either. You could introduce a new vendor NPC that sells literally every obtainable object in the game in searchable format, without any sadistic nonsense like randomized inventory or blind gambling. You can even make the prices exorbitantly expensive, to ensure that players still have to earn them – as long as they can work directly towards collecting a full set of armor instead of needing to rely on dumb luck.

In fact, randomizing vendor stock is an undesirable custom regardless. If i need a Draining Touch III spell for my build, i’ll exit and enter town however many times it takes to buy it from the magic vendor NPC. Randomizing her inventory doesn’t prevent me from obtaining whichever item i want. All it does is make the transaction unpleasant. Resolving this issue can’t be that difficult. Basically it boils down to one extra user interface design task.

Another feature synonymous with open-ended RPGs is dungeon layout randomization, which usually produces multiple dead ends. Of course there’s nothing inherently wrong with closed paths. Map exploration would feel far more linear without the occasional wrong turn.

The boring part is walking all the way back to the fork in the road. Torchlight sporadically addresses this backtracking tedium with a sort of trap chest which spawns an undead mob to repopulate the return trail. It’s a great idea, although it doesn’t happen often enough.

What about adding an intra-level portal spell to create instant waypoints controlled by the player? Navigation efficiency would be greatly enhanced by such a tool. For example, one portal could be placed wherever the main path branches off. Upon reaching a sealed alleyway, the next portal would transport the player back to the previous one’s location.

To prevent players from exploiting portals as an escape hatch, simply give the spell a painfully slow casting period – around one or two full seconds. That should be enough to get them killed for trying to activate one in the middle of battle. (If the dual-portal concept seems too similar to Valve’s Portal, the same functionality can be implemented through teleport beacons or destination markings to summon transport creatures.)

Speaking of teleportation, RPGs typically prevent players from phasing across barriers or chasms – which is perfectly understandable. Level designers may want to preview later parts of the dungeon or hint at secret locations without allowing direct access.

The question is, why can’t players teleport through those obstacles in the opposite direction? Couldn’t the game check the automap to see if they’ve cleared a path between the two locations? If so, allow them to teleport across. If not, then restrain the teleport to force them to traverse the intended path at least once.

Last but not least, please fix your combat! If my mouse is clearly hovering over an immobile opponent, with his lifebar displayed on screen to indicate target lock, my arrows should not be missing him from two meters away. Sadly, this happens all the time in Torchlight. Moreover, if my magic spells instantly dissolve into thin air whenever i’m standing on a staircase or near a wall, then the game’s projectile collision system needs work.

If necessary, hire a talented combat designer alongside a veteran gameplay programmer with experience working on fighting games; but make sure the most basic interactions in the game are intuitive and responsive. Combat design is too essential to the core gameplay experience of an action RPG not to make it a priority.

Including a proper training mode couldn’t hurt either. It’s been a great addition to fighting game development, and even action franchises like God of War and Devil May Cry have begun experimenting with spinoffs and variations on the concept. Who wouldn’t enjoy testing out different skill tree configurations before committing to a hardcore ladder build? Should character exploration be so time-consuming?

Granted, traditional RPG producers might resist the notion of giving players immediate access to all skills and items, even in a limited practice mode environment. Obviously some online RPGs are designed to keep players hooked on acquiring incremental rewards.

Well, how about providing only as many level bonuses, skill points, stat points, and items in training mode as the player earns through standard progression? Venturing further into the main campaign would unlock more upgrades in training mode and repeal level requirements. That seems like a reasonable compromise, right? At least that way, players won’t have to grind all the way up to lvl50 nine times to explore three classes with three skill trees.

Hopefully some of these suggestions sound interesting enough for someone out there to try them. If you happen to be that person, good luck – and thanks.

Now that i’ve seen almost everything there is to see in Torchlight, i have to say that i’m extremely disappointed with how unnecessarily long it took to reach this point. Don’t get me wrong, the game is expertly crafted in many ways, but my lasting impression of the experience is frustration.

As much as i want to show off screenshots of Torchlight’s compelling atmosphere, i can’t bring myself to recommend the game to anyone – for the simple reason that it quickly turns into addictive torture by design.

Believe it or not, the whole reason i played the demo was because the XBLA description promised that i could transform my pet into various creatures, including elementals. I’ve had a soft spot for elementals since the original Warcraft.

What that marketing blurb didn’t explain is that you have to “play” their braindead fishing “mini-game” for 5-10 hours straight in order to catch one extremely rare fish to permanently transform a single character’s pet into a (random!) elemental. I mean, why??

Can you imagine how many productive hours have been utterly wasted in search of that tiny speck of satisfaction? It’s not even a challenge. It’s just dumb, pointless grinding. How could Runic Games get so many things right, yet still relapse into that old sadistic routine, which sadly still permeates the open-ended RPG genre?

The truth is, Torchlight is a great game until your character reaches lvl40. Beyond that point, it has nothing of substance to offer – but it’s terrified to admit that. As long as you stop at lvl40, you’ll remember it fondly. If you keep going, what little you find won’t justify the growing time investment.

If you want to continue playing, you’re far better off starting a new character. All three classes are interesting in their own way, so there’s no wrong choice between them. Instead of replaying the main campaign, you might want to try heading directly to the Shadow Vault in the southeast corner of town.

Maybe the coolest part of taking the Shadow Vault route is that you never have to lose Brink. Normally he accompanies you from the second level of the main campaign until the first boss battle on the fourth floor. As long as you avoid that fourth level, he’ll never leave your side. Brink isn’t exactly dominant in later stages, but he’s free and it’s still fun having him around.