On a rooftop in Gotham City in the video game Batman Arkham Knight, I had to beat up six ninjas without getting hit. It took me 46 attempts, one after the other, during one frustrating hour.
This article originally appeared in October, 2015.
It wasn’t the easiest thing I’ve done in video games, nor was it the hardest. It was, however, the first time I ever recorded this kind of serial failure and eventual success.
I needed to see what happens when I try and try again.
Looking over the footage, which you can see below, I learned a few things. These lessons might only apply to how I handle rooftop battles in Batman games but may apply to other things in life, too:
- Faced with a difficult task—say, KOing six virtual ninjas without getting hit—I’ll rigorously try the same failing tactic too many times in a row. I’m so sure it’ll work the next time.
- Near-success doesn’t breed immediate successive success. If I do really well but still fall short at a challenging task, I’ll badly flub the next attempt.
- I don’t want to take breaks while repeatedly trying something difficult, but if I do and then return, I do better.
- Any sound effects meant to encourage me when I’m doing something well make me nervous when I hear them. They’re the sound of pressure.
- Gradually, if I keep trying, I will do better and I will eventually succeed.
Again, maybe these only apply to playing Batman games. Maybe not! See for yourself:
What you see below is what all that looks like when I’m bashing my head against one challenge in a Batman game.
Remember, I don’t want the ninjas to hit me. Oh, and I’m not playing as Batman. I’m playing as a guy called Azrael.
- Attempt 1 (five seconds) - I’m hit right away by a swiping blade attack.
- Attempt 2 (four seconds) - I’m hit right away by aerial ninja attack (why didn’t I block???)
- Attempt 3 (nine seconds) - A meek attempt by me to snare a ninja with a grappling hook attack. No good. I quickly fend off a double attack but am hit in the scrum by an ordinary attacker.
- Attempt 4 (six seconds) - Another grappling hook attempt, still not helping, quickly followed by an explosive gel attack, then a successful defense against an aerial ninja attack (signs of improvement!) but then I get hit. Failure.
Batman Arkham Knight lets you retry challenges very quickly, which helps the player stay in a state of psychological flow, according to Jamie Madigan, a psychologist, gamer and author of a book about gamer psychology called “Getting Gamers.”
I’d told him about my many attempts at the one challenge in Arkham Knight, a challenge he knew well and that he figured would get me in that state of flow. “That’s the ‘in the zone’ feeling you get when the challenge of a task is balanced well against your ability,” he explained.
“If you’re getting closer and closer to beating the challenge over time and you’re getting clear, immediate feedback about what you’re doing wrong, then you’re likely to experience a state of flow. The Batman combat systems are good at this because it’s usually clear what went wrong based on on-screen prompts: you didn’t do the double counter right then, you broke your combo right there, you got shot by a thug with a gun right there, you tried to punch a guy with a shield right there, etc.”
- Attempt 5 (12 seconds) - I’m still starting with gadget attacks. I get surrounded by three enemies, which should lead me to spam triangle and do a series of counters. Instead I flip over an enemy and try to isolate him. Soon enough, I’m overwhelmed.
- Attempt 6 (five seconds) 1:13- I finally do a good multi-enemy counter, but a third enemy approaches with the blade attack that did me in during my first attempt. Those blade attacks are signaled by some jagged yellow lines that appear above the attacker’s head. I should be able to do a special block, but instead I just get sliced.
- Attempt 7 (four seconds) - One good counter and then sloppiness.
- Attempt 8 (five seconds) - Aerial ninja attack blocked—it’s no longer a problem—but I’m stymied by a swiping blade attack again.
- Attempt 9 (four seconds) - I’m back to starting with gadget attacks. Total failure. Aerial ninja attack does me in.
- Attempt 10 (12 seconds) - 2:00. I start with flailing gadget strikes, but maybe I’m on to something. It’s sort of working? Then some group counters. Yellow blade attack starts and I block the first two swipes successfully. I get cut on the third. Progress.
- Attempt 11 (19 seconds) - I’m still starting with the flailing gadget strikes, then some good counters. My longest run yet! But the blade attack does me in yet again. The game must know I have a problem, because it is now giving me a prompt advising me how to counter the blade strikes correctly.
- Attempt 12 (8 seconds) - Some text shows up on screen telling me how to deal with the blade attack, but I’m busy blocking and countering standard attacks. An enemy does the blade attack. I get cut on the first swing. The correct action involves a button press while pulling back with one of the controller’s analog sticks. From the way my character moves, it looks like I was pulling back, but my timing was off.
- Attempt 13 (five seconds) - A few successful moves, but the first blade attack gets me.
- Attempt 14 (four seconds) - I’m falling apart. The very first enemy attack is an aerial ninja attack and it gets me.
- Attempt 15 (22 seconds) - 3:22. Turnaround! I successfully counter a group of three enemies. And, even better, I finally dodge all the swipes of a blade attack. But a regular attack gets me.
- Attempt 16 (eight seconds) - Sloppy. Felled by a regular attack.
- Attempt 17 (17 seconds) - A solid run. I’m mostly waiting to counter. Blade attack successfully dodged. I’m done in by a regular attack.
- Attempt 18 (five seconds) - I get more aggressive but I also get hit quickly.
The longer runs lead to quick stumbles. I see this as some sort of failure to immediately build on progress, but Madigan thinks I’m underestimating the role of chance. “It often feels like progress is followed by failure, because there is some randomness in your performance,” he said. “If, by luck, you have one really good run you’re unlikely to do as well or better on the next, because your next run is most likely to be average or close to it. That’s just how the normal distribution (that is, the bell curve) works. Most of your attempts are going to be clumped around the middle, with only a few breaking away to the extremes on either side. But over time your average attempt should improve because of an increase in skill. We just don’t normally think about stuff like this in terms of a distribution of scores.”
He’s right in that I kept hitting higher peaks. I was, in a general sense, improving.
- Attempt 19 (24 seconds) - 4:42. Easily my best run. Lots of counters. I’m getting more aggressive. Done in by a regular attack.
- Attempt 20 (18 seconds) - 5:13. A really interesting run, in retrospect. I do aggressive counters and am clearly getting faster and responding to enemies. I hit the first blade attacker before he can even take a swing. I dodge four swipes from a second blade attacker, but I am clearly at a loss about what to do to him after the fourth dodge. While I stumble trying to figure that out, a regular attack hits me.
- Attempt 21 (10 seconds) - It starts well but ends quickly. Sloppy.
Six minutes in, I leave the game idling.
I resume playing about a minute later.
- Attempt 22 (15 seconds) - I do well after my short break, including a successful four-swipe blade dodge, but I’m once again at a loss about what to do right after that and get hit by another enemy’s regular attack—right after those dodges.
- Attempt 23 (five seconds) - A shorter run. The Batman voiceover says “stay focused” but I think I’m losing that focus.
- Attempt 24 (five seconds) - An older ploy returns! I start with a gadget attack for the first time in a while but am quickly felled by an enemy’s regular attack.
I’d lamented to Madigan that I kept going back to earlier strategies that had already failed and he convincingly explained to me why I’d be going back to stuff that I should have known was still not going to work. “Regression to the mean can trip us up because of what’s known as the ‘post hoc ergo propter hoc’ fallacy—Latin for ‘after this, therefore because of this,’” he said. “Basically ‘X happened after I did Y, so Y must cause X.’ [That] is not always true. If we try something new—say use the batarang or the batclaw more in combos— it may result in a better run. But when our next run is closer to the average (i.e., worse) because of random chance, we may think that the new strategy isn’t working after all or even that it’s causing worse performance.”
- Attempt 25 (five seconds) - An unusual enemy attack pattern starts things off and confuses me. I get hit with the second swipe of a blade attack.
- Attempt 26 (four seconds) - 8:19. For the first time since I started, the first move from my enemies is a blade attack. It gets me.
- Attempt 27 (six seconds) - I make a few successful counters and then get hit by the second swipe of a blade attack. Whatever progress I was making has decayed.
I take another break and resume about two minutes later.
- Attempt 28 (17 seconds) - By my recent standards, this is a pretty good run. But I still get his pretty early. I am hitting blade attackers before they swipe, at least.
- Attempt 29 (nine seconds) - A few good counters, but I’m hit by a blade attack.
- Attempt 30 (five seconds) - I start with a gadget attack again (explosive gel) and do a couple of counters, but I’m hit by a blade attack.
- Attempt 31 (16 seconds) - 11:28. The computer starts with a blade attack and I handle it effectively. I counter and dodge enemies at will and am doing well until I get hit with the fourth swipe of a third blade attack. It is as if the artificial intelligence knowns my weakness and keeps spamming the same attack on me.
- Attempt 32 (six seconds) - A weird one in that it seems that I do counter the guy who hits me. Not a great run.
- Attempt 33 (eight seconds) - There’s a nice counter early but also some sloppy play. I get hit fast.
- Attempt 34 (22 seconds) - 12:21. My best run yet and possibly the first in which I finally start knocking enemies out. A loud noise sounds when I get a KO, and I think it distracts me. I get hit right after I hear it. I got too excited!
- Attempt 35 (six seconds) - A short, bad run. I’m still not following good runs with good runs.
- Attempt 36 (nine seconds) - Nothing special. Hit with a blade attack.
- Attempt 37 (13 seconds) - A more aggressive run than usual, and I score a knockout, but once again, as soon as I hear the KO sound, I get hit with the next attack. I’m getting rattled by my moments of success.
- Attempt 38 (18 seconds) - Two knockouts and, even better, I don’t fall apart as soon as I hear the KO noise. I hold my own for a few more seconds before getting hit. That’s two pretty good runs back to back, for the first time.
- Attempt 39 - (five seconds) 14:07. A terrible short run. My streak of good runs is over at two.
- Attempt 40 (24 seconds) - Strong run with three knockouts and even some good handling of blade attacks, but I clearly get a little lost and slashed with a blade attack.
I take a break to have dinner. I should know that this kind of break helps, but it really is hard to put a controller down and take a break. I do most of my gaming on the weekends, especially if I’m playing a game for review, and I’ve found that I usually finish tough games on Sundays. I’ll struggle mightily and repeatedly to beat the difficult final section of a game late into a Saturday night, then go to bed, wake up and clear it in my first or second morning try.
Dinner’s over, and I try again.
- Attempt 41 (14 seconds) - I score one KO but a lot of flailing. A so-so run.
- Attempt 42 (31 seconds) - This is a confusing run that has a few KOs! That dinner break paid off. But I get taken out with a blade attack.
- Attempt 43 (13 seconds) - Some countering at first but I get hit by a blade attack.
- Attempt 44 (8 seconds) - 0:26. Garbage run. After a couple of counters, I’m hit with a blade attack. I’ve once again lost my ability to deal with blade attacks. Confidence failing.
- Attempt 45 (10 seconds) - Some improvement and one KO, but I get hit shortly after the KO noise sounds. It’s still getting me too excited.
- Attempt 46 (50 seconds) - 2:00. Complete success. I fend off seven blade attacks and knock out all six enemies. It’s not pretty. I wind up walking around a few times mid-fight, but clearly my sudden ability to handle any and all blade attacks is my key to victory. I score four KOs in a row to finish. Credit learning to finally stay calm when the KO noise sounds and finally learning the timing of defending against the blade attack. Batman seems pleased, and so am I.
Obviously, Batman Arkham Knight is just one example—I’ve tackled challenges like that in video games thousands of times. Everyone who has played games has. We want to get through this stuff. I think we even enjoy getting through this stuff.
“We keep attempting challenges like these simply because they act as a challenging goal,” Madigan said. “Research on goal setting has shown that having the goal be explicit and detailed anchors our expectations about our own performance. [We think:] ‘This challenge is in the game, it must be possible. I must be able to do it if I keep trying.’”
We can all probably learn something from video games about how we deal with challenge and failure. Games are laboratories for failure, gyms that generate eventual success. You may play a game to have fun, but you can also play it knowing that it’s a part of your life that you may initially stink at but will eventually master.
Sure, you may never win Tetris, but you’ll last longer the more you play.
You’ll only fall in Super Mario’s bottomless pits for so long before you always remember to jump.
The cops will get you in GTA, other players will snipe you in Call of Duty, but you will improve.
Games, unlike the rest of life, are fair that way. They reward practice and dedication. Their trials reliably lead to treasure.
Eventually we learn how to dodge the swinging blade attacks. Eventually we realize that starting with a gadget maneuver isn’t going to help. And on the 46th try, after a brief dinner break, we succeed.
Bonus: If you really want to watch all 46 attempts, here you go...