I’m not one for New Year’s resolutions. The concept is simultaneously noble and flawed, brilliant and defective. As humans, we know we are capable of (and require) constant growth and improvement. We rarely make a conscious effort to change, though. That’s why January 1 is such a tantalizing reset button. This iteration is perhaps even more compelling, as 2020 marks not only a new year but a new decade. A new year is a new chapter in our lives. So why not start that reading plan or workout routine and become a brand new person by next December? We’re all familiar with what happens next. Work is hectic on Thursday and by the time you get home, you’re too exhausted to read anything. Or maybe there’s a big school project due next Friday and honestly, who has the time to work out? And the worst of all, you wake up in the morning and for some reason, just don’t feel like doing what you promised yourself you would do.
New Year’s resolutions are rarely maintained and frequently broken. They don’t seem to stick for most people, myself included. That’s why I’m not particularly fond of them. I am, however, all for setting good goals and reaching them. So what separates someone who makes a New Year’s resolution to run more from someone who actually completes a marathon? I’d like to find out. In a previous post, I documented my cubing journey over the last three years. I’ve made steady progress, but I know I can do better.
Where I Am
Unofficially, my best single is 10.46 seconds and my best average of five (Ao5) is 13.68. While I trust the validity of those numbers and use them when people ask what my best times are, those are times at home. The World Cube Association (WCA) doesn’t recognize these times for obvious reasons.
This month, I participated in my first WCA competition in almost 2 years. My previous best results were abysmal: a 17.98 single and a 21.77 average of five. But as I mentioned before, I have been practicing and seeing steady gains(unofficially at home). It was time to formally prove I was as good as I thought I was. And the results are in: I’m officially sub-20! I finished with an Ao5 of 17.43 seconds and even managed to sneak in two sub-15 singles. (Ignore the 38.74 solve in the first round; I messed up.) My new average time is faster than my previous single time. That’s amazing! I was thrilled with these numbers. This is my official benchmark at the end of the decade: a sub-15 single and a sub-20 average.
Where I’m Going
One of the biggest challenges to improvement is a lack of clear, defined and achievable goals. If you have no destination, you’re just wandering. I’d prefer not to waste my time like that, so I’ve got something in mind. For as long as I can remember, I’ve had two goals in cubing:
- Maintain a sub-15 average
- Achieve a sub-10 single
Historically, the top 100 averages in the world were all sub-15 in 2008, and the top 100 singles were all sub-10 in 2010. I’m aiming for that level. (As of the time of writing, the top 1000 singles in the world were all under 7.5 seconds.) I’ve occasionally joked that once I hit those milestones, I’ll retire from cubing entirely. So how do I shave two-and-a-half seconds off my average and five seconds off my single time?
How I’m Getting There
There’s a concept known as deliberate practice. It’s a type of practice that goes beyond rote memorization or repetition, focusing on iterating and improving the technique. There are scores of resources designed to help people improve their skills in a variety of areas: running, piano, baseball, and violin, to name a few. There are many tools (many good ones, at that) for cubers. There are hundreds of videos explaining specific algorithms and concepts, walkthrough solves, and more. But I can watch 40 videos on F2L lookahead and never improve a bit. So why not take concepts that have been developed over decades for other fields and apply them to mine? I’m going to discuss a few big ideas that have been shown to work elsewhere, then examine some ways I can specifically implement them for my purposes.
Before I begin, a caveat: these are techniques I’m customizing for where I am in my cubing journey. They may or may not apply to you. If you’d like to enter this wild and wonderful world, I’d recommend J Perm’s tutorial for solving a cube. If you’re sub-60 and want to get faster, Max Deutsch documented his journey from there to sub-20 in a month. If you’re under 20 seconds, you probably already know where all the good resources are. If not, YouTube is a great launching point (try looking for J Perm, BrodytheCuber, or cyotheking, among many others). Let’s get started!
Set Specific Goals
The goals I have are nice, but Dr. Ericsson says that I need to be more specific:
If you’re a weekend golfer and you want to decrease your handicap by five strokes, that’s fine for an overall purpose, but it is not a well-defined, specific goal that can be used effectively for your practice. Break it down and make a plan: What exactly do you need to do to slice five strokes off your handicap? One goal might be to increase the number of drives landing in the fairway. That’s a reasonably specific goal, but you need to break it down even more: What exactly will you do to increase the number of successful drives?
Reaching a sub-15 average and sub-10 single are good final goals. But what do I have to do to get there? I don’t have specific benchmarks for my times on each individual step of the solve, so these numbers were just taken from this guide in step 6 and other steps adapted for my case.
- Cross
- All cases under 8 moves
- Finish under 2 seconds
- F2L
- Improve lookahead
- Insert 2 pairs at a time blindfolded (untimed)
- Slow solves
- Finish under 8 seconds
- X-Cross (?)
- Plan first pair in inspection (?)
- OLL
- Finish learning all OLL cases
- PLL
- Finish under 3 seconds
I think my cross is fairly efficient, if not very fast. I haven’t actually measured myself, so I couldn’t say for certain. The section that stands to reap the greatest reward is F2L, without a doubt. OLL and PLL can only shave a couple seconds off my time, but improving my lookahead (and thus my F2L) will have the biggest impact. I’ve seen multiple sources suggest two pair blindfold practice to be effective, and it makes sense. The whole point of lookahead is to prepare to do one thing while doing something else. I still need to measure my OLL and PLL times, however. My sister recently pointed out that my algorithm executions are almost rhythmic. Great for music, but terrible when pure speed is the goal.
Don’t Start at the Beginning
When practicing, I generally just scramble, solve, and repeat. This is fun and useful to an extent, but doesn’t result in great improvements. Instead of sitting down for a 50-solve session, practice only F2L or PLL one day.
Practice Without Practicing
When not actively “practicing”, it’s still possible to passively get better. The size of the cube lends itself very easily to this. Walking to class is a good time to practice a few algorithms. I sometimes do this while walking to class, but it’s mostly composed of me fiddling with my cube aimlessly or infinitely repeating T Perms. If I plan out what specific algorithm I need to drill on a a given day, it would be very easy to just grab the cube and go.
Write It Down
Humans are terrible at remembering. I’ve been a strong proponent of journaling for years and have never thought to apply it to my practice routines. I don’t think an incredibly detailed journal would be useful in this case; perhaps just little notes as I notice them during a practice session.
I think a one implementation would be as follows: After a solve, rate each step as good or bad. There are four steps to deal with here, so a five-point scale for each step may be somewhat overkill. So cross may be good, followed by a terrible F2L, then good OLL and PLL. Every once in a while, writing down what was good or bad about the solve may be useful. This is a rough draft of an idea, completely undeveloped, but the premise is there. This may be useful when practicing individual sections such as only F2L or only OLL. It may be somewhat a bit much to rate four sections of a solve every time.
Final Thoughts
By December 31, 2020, I would love to see a sub-15 average and a sub-10 single next to my name. In the past, I would solve and solve and solve, hoping to end up with an occasional sub-15 or perhaps a 10.xx second time. Now, I have a plan to get there. I intend to check back in here occasionally and document my progress (what’s the point of the destination if you can’t enjoy the journey?).
The future is exciting. The point of this exercise is to inspire hope beyond just solving cubes. If there’s something you’d like to see change in your life, why not make a plan now? It may not go exactly as you imagine, but nothing will happen if you don’t take the first step. “The future marches in and calls us free.” What are you going to do with it?
Update: January 31, 2020
It’s been a month! School has been incredibly busy, so I haven’t had much time to cube since the semester started. I did manage to benchmark my Cross + F2L time. I averaged 16.45s over 131 solves. This is odd. Over the last month, I’ve had 165 solves averaging 18.57s. That would imply that on average, my OLL and PLL take two seconds. On a gut level, this seems wrong. It may be correct – I haven’t benchmarked my OLL and PLL – but instinctively, I think I take longer. On the plus side, at least my overall average is decent? My official average is more than a full second faster though, so we still have a lot of work to do.
Update: February 29, 2020
Happy leap day! I managed to complete an average of 100 solves this month. It didn’t happen all in one session; I took a few breaks in between to stretch, wash dishes, get food, etc. But my average was 17.03 seconds! That is absolutely fantastic in my book. What’s remarkable is that I haven’t spent a lot of time in the practice sessions that I detailed previously. I did notice two things from this session, however. 1) My lookahead is much better! This is such good news. Previously, I would try not to focus on my current pair and just hope to see something else useful. Now, I’m actively tracking pieces, rotating to eliminate blindspots, etc. It’s much smoother. 2) I had a couple LL skips that resulted in 13.xx times. This tells me that my Cross + F2L time is somewhere in the ballpark of 14 seconds. This make far more sense compared to my previous estimate of 16 seconds. Of course, the 13 second solves were very good F2L solutions, not necessarily my average. But it’s still reasonably close. Overall, I’m very happy with how this month turned out! Not nearly as many solves as I had hoped, but very good progress. Quality beats quantity!
Update: November 25, 2020
It’s been a…long time! As you can see by the date of the last update, the world had not yet drastically changed. Combine all that chaos with my final semester of school and moving on into the “real world”, these updates fell by the wayside. The cubing, however, did not. I did not think that my times had improved significantly over the past summer and fall. In fact, if you look at a graph of all my solve times, they seem to be holding steady over the last several months. But if you look at the share of sub-x times, you see a completely different story. See below:

This graph is an updated version of the original analysis I did back in October 2019; you can read that post here. The data that immediately caught my eye is the sub-15 percentage. In early 2020 (January – April), just under 7% of my solves were sub-15. That number dropped to an abysmal 2.5% in May, likely due to the fact that I had fewer than 200 solves that month; I had been averaging over 400 monthly through the first four months of the year. My sub-15s rebounded in the summer months of June and July, but they absolutely skyrocketed beginning in August. I don’t think I consciously noticed the change, but the numbers are plain as day. I have almost doubled my sub-15 times to comprise ~13.4% of my solves now. In fact, I was averaging over 500 solves per month from August to present day. So my raw numbers are increasing, and I’m seeing an uptick in faster times. That’s great! My most recent average of 50 solves clocked in at a cool 16.55s with over 20% of those solves under 15 seconds. Counting just this month’s solves (679 of them), I am now sub-17 with a borderline 16.99s average. With that result, I just need to lose two seconds to get to my ultimate goal of consistent sub-15 solves.
The irony in all of this is that I don’t feel any faster; in fact, I’m catching myself making more mistakes than ever. Perhaps that’s a sign of growing pains. I’ve also noticed that my mentality recently has been focused on breaking the elusive sound barrier that is the 10-second solve. I wonder how much that has contributed to the overall improvement. At any rate, this data is an encouraging milestone and I’m going to celebrate it while I can. Tomorrow is a new day to get that much closer to the dream of a sub-15 average and a sub-10 single!