• You are not logged in.

    Practice with alt/floating/same-finger bigram combos

    • Started by MicahElliott
    • 1 Replies:
    • Reputation: 0
    • Registered: 11-Jun-2020
    • Posts: 2

    I've been thinking/reading (https://forum.colemak.com/post/8851/#p8851) about same-finger inefficiencies and so enumerated which bigrams seem common/troublesome. Thought I'd share what I just started doing to recognize and get better with these....

    This is not a scientific analysis, just a simple word count from OED. I'm sure it would be better to get word usage frequencies, but I don't think it matters for just practice purposes.

    The bigrams (with rounded OED appearance counts):

    - NK/KN (380)
    - LK/KL (210)
    - DG (130)
    - MN/NM (100)

    - SC (720) **
    - UE/EU (500)
    - WR (120)
    - NH/HN (120)

    The first group is amenable to finger "rearrangements" (and then you get a few really interesting trigrams to figure out where you're shifting 3 fingers: ULK, UKN, etc). The second group might need some sort of "sliding" (UE, WR) (and maybe assuming you have a matrix layout). The SC is a special case that doesn't seem slideable since you're so curled, but a middle-then-index is feeling comfortable enough with it.

    If you have bash or something like it, you can generate lists of these to feed into https://www.keybr.com/ (see Settings -> Provide your own custom text)

    ❯ echo `cut -f1 -d' ' $dict G kn |tr '[A-Z]' '[a-z]' | shuf`  # $dict is some path to where you keep your dictionary file, 'kn' is the bigram you wish to practice
    knoll hackney jackknife knitting unknown well-known clasp-knife knowingly thickness knee-jerk knowledgeable flick-knife paper-knife hackneyed knelt foreknow knave sickness unknot knuckleduster knapweed knight kneel knickers knee know-how hackney knife nickname knew fish-knife knit reef-knot unknown palette-knife knowledge knocking-shop banknote knacker knobbly knack unknowable unknown unacknowledged knell weak-kneed knead knurl knuckle knickerbocker knot1 knockout knotty unbeknown knuckle close-knit knapsack knee-bend knock knife-pleat acknowledge knock-on self-knowledge kneeler unknown kneecap knight knot-hole knotgrass knob knee-deep acknowledgement cockney cracknel knockabout topknot knick-knack knocker known penknife knives knee-high slip-knot knot2 weakness knickerbocker knight breakneck knee-length know knock knitwear know-all knowing knock-up knees-up unknowing knitting-machine antiknock knock-down camiknickers knitting-needle knee-breeches knife-edge doorknob

    Maybe it's only worth focusing on LK, NK.

    For "down-sliding" I'm considering if it's worth getting shorter caps for U (to get an easier UE), etc.

    Feedback/ideas appreciated. I could have missed some good combos, and I can explain more if what I'm describing isn't clear.

    Last edited by MicahElliott (04-Nov-2020 21:02:52)
    • 0
    • Reputation: 0
    • Registered: 11-Jun-2020
    • Posts: 2

    I'm guessing it's already well known to this audience how frequently bigrams show up, but for anyone curious about the set of somewhat common same-finger bigrams, here's a rough and ugly (and maybe totally wrong) analysis, using this word frequency data set: https://www.kaggle.com/rtatman/english-word-frequency

    It just adds up the frequencies of the words in each bigram set.

    ❯ for bg in nk kn sc lk kl dg pt ue eu wr mn nm dv fs yi hn hm nj ; do freq=$( for w in `cut -f1 -d' ' $dict G $bg |tr '[A-Z]' '[a-z]' `; do grep "^$w," unigram_freq.csv; done | cut -f2 -d, | paste -sd+ | bc); print $freq $bg ; done |sort -n
    26871090 fs
    114257917 kl
    124769039 mn
    144033872 hm
    148160741 nj
    382106238 yi
    437057914 eu
    455586061 nm
    467764973 dg
    602509305 dv
    621954800 wr
    657595295 lk
    859167332 kn
    1119200013 hn
    1698212931 nk
    1865938538 pt
    3497019606 sc
    4162567648 ue
    • 0