My typing speed has been historically poor. Naturally, I decided to track my performance on a spreadsheet. After three months of recording my results from completing typing tests (amounting to over 9 hours of literally typing), here is some commentary on my process:
Methodology
I started with Typing Academy to practice typing, but I also later moved on to other websites, namely Monkeytype and TypeRacer. Typing Academy tests using a random string of the most common words in the English language, and it was the only website I used that doesn’t allow typos to be deleted. While this meant I could kind of inflate my words-per-minute (i.e. WPM) metric, losing concern for using the backspace key, this also made the tests less “realistic.” Monketype also uses a series of random common words, but has a much more pleasing user interface (in my opinion). It tracks more metrics for each test, and incorporates the usage of the backspace into the test. TypeRacer, in contrast, grabs from a database of user-submitted excerpts from various texts on the internet. With sources ranging from Bruno Mars to J. R. R. Tolkien, these texts also required capitalization and punctuation, thereby giving a more “real-world” guage of my WPM. While Monkeytype was the most “fun” to use, I’d say TypeRacer improved my typing the most.
To document my results, I used Google Sheets. I wasn’t really sure at first what insights I wanted to get from my tests, so I mostly recorded for variables I thought could be helpful. For each 60-second typing test I completed on whichever platform, I logged the WPM, the error rate, the date, which keyboard I was using (my laptop’s or an external membrane keyboard), and which site I took the test on.
Visualization
I also used Google Sheets to create the graphs. The main one I look at plots the daily average WPM for each website over the day:
This graph automatically updates as I enter in more data points.
Analysis
Most notable is the consistent upward trend in the first couple months of each website. While I’d like to attribute this to my dedication in improving my typing skills, I also acknowledge the confounding variable of having seen the same words before. By the same token, the hiatus during spring break and afterwards resulted in noticeable plateaus to my WPM, an effect of being “rusty.”
Conclusions
The following are some takeaways I’ve gathered from this ongoing activity, some of which seem to fit into broader ideas:
I found that the laptop keyboard is more error-prone than a membrane keyboard (since there’s less travel distance), but this also means that laptops can be faster and less fatiguing.
Errors and speeding were more costly to WPM than a more conservative pace. It’s worth noting however, that spell check can batch-correct errors. For practical situations, you are probably better off blitzing and spell-correcting. Playing to maximize the metric of WPM would defeat its purpose.
Making progress visible gamifies the process of improvement, making the overall process of practice typing more enjoyable.
Great blog, Kevin! I think it’s really interesting how you’re trying to optimize a lot of basic things in life that many would overlook or take for granted. It’s obvious that you’ve really taken the time to reflect on where your time is wasted and how to improve.
One thing I’m unsure of is how the payoff in efficiency from these minute improvements compares against the relatively steep learning curve to adjust habits. I think it’d be interesting if you could perform another study to determine the savings that a simple keyboard shortcut could accumulate over a period of a day, a month, or even a year.
Another interesting thing to consider is how these optimizations apply to organizations. This seems like a vastly more difficult undertaking because an inverse relationship exists between size and efficiency. Obviously, the free-market structure makes firms more competitive and efficient, but are there other incentives or modes of organization that would do better?
Hi Kevin,
I must say, the scientific nature of the title of your blog was what first attracted me to your blog. Needless to say, I very much enjoyed reading your blog. The fact that you documented and explained almost every part of the process like a scientific experiment was very impressive. The data was especially interesting to look at, and it reinforces an important concept, namely progress is gradual. So often we search for instant gratification in the form of instant improvement, but more than often this is a pipe dream and is rather unrealistic. The documentation of your progress is an important reminder of the graduality of progress. Your data especially emphasizes this point well, where at certain points the typing speed actually decreased. However, over the long-term (multiple months), we are able to see how you have made significant progress. Making these small improvements over a long period of time tends to add up, and I can say at a personal level you have inspired me to track my own personal growth. I’ve always had a goal to get a minimum of 7 hours of sleep a night at around the same time, and I think mirroring your data collection process will help me to stay more accountable to myself. Great blog, and keep up the good work.
Wonderful Kevin. I am so glad, that after months of tracking, some competition, some yoking, some banter, we see the journey of Kevin’s typing. Characteristic to your work, this analysis is quite detailed and thorough, and it makes it very clear as a reader what factors played key roles in typing speed.
I am a fan of Monkeytype because of the interface as you mention. For some reason, the way the letters load into the screen and the words progress just makes my fingers fly faster. But then again, Monkeytype’s default settings automatically set to 30 second bouts, meaning this is a anaerobic exercise, a sprint if you will. Typing is mostly a marathon, an endurance type of activity in my opinion, so perhaps typeracer is more practical.
Regardless, I love how you include all these other factors too, like keyboard usage. I forget if you’ve ever gotten a mechanical keyboard, but man oh man do those things go. When my fingers are on those perfectly pressurized keys, I’m in heaven. The school desktop keyboards are honestly pretty good too, but I say that the Mac keyboards are arguably one of the bets, with the low button press depth, allowing for faster touch and release form, if you will. I personally despise my current laptop’s keyboard because it’s simply broken. It frequently double prints the letter s or a or o or literally anything at any point for not apparent reason and it makes type testing quite frustrating.
Anyways, I’m glad to see the trend going roughly upward. Next, you need to conduct a linear regression test to find a linear model for training time and typing speed