Stats table comparing alt keyboard layouts
Pascal Getreuer, 2023-11-27 (updated 2025-08-22)
This table expands on the one in my my alt keyboards post, comparing English typing stats for various alternative keyboard layouts.
Lower is better for SFBs, SFSs, LSBs, scissors, redirects, and off-home pinky use; higher is better for rolls. Explanation of the metrics can be found in the Keyboard layouts doc. Metrics were computed for English with Cyanophage’s layout playground.
| Layout | SFBs | SFSs | LSBs | Scissor | Rolls | Redir | Pinky off |
|---|---|---|---|---|---|---|---|
| QWERTY (1873) | 4.38% | 5.45% | 4.55% | 1.46% | 40.76% | 6.22% | 2.47% |
| Dvorak (1936) | 1.87% | 3.48% | 0.80% | 0.08% | 39.20% | 1.55% | 4.13% |
| Arensito (2001) | 0.96% | 3.72% | 1.22% | 0.45% | 54.55% | 5.28% | 1.48% |
| Capewell (2005) | 1.19% | 3.34% | 0.62% | 0.19% | 51.43% | 7.89% | 1.64% |
| Colemak (2006) | 0.91% | 4.24% | 2.26% | 0.26% | 49.20% | 5.33% | 0.78% |
| Workman (2010) | 1.97% | 4.15% | 1.11% | 0.47% | 47.40% | 6.05% | 0.78% |
| MTGAP (2010) | 0.92% | 3.27% | 0.46% | 0.15% | 46.34% | 1.78% | 3.80% |
| Colemak-DH (2014) | 0.91% | 4.24% | 1.27% | 0.15% | 49.20% | 5.33% | 0.78% |
| Hieamtsrn (2014) | 0.79% | 3.56% | 0.56% | 0.11% | 44.63% | 1.49% | 4.18% |
| Halmak (2016) | 1.97% | 3.47% | 0.40% | 0.56% | 40.11% | 2.52% | 5.48% |
| BEAKL19bis (2020) | 0.95% | 4.07% | 1.52% | 0.39% | 43.74% | 1.50% | 2.26% |
| Handsdown Neu (2021) | 0.76% | 3.99% | 1.26% | 0.42% | 44.04% | 1.47% | 2.89% |
| Boo (2021) | 0.78% | 3.03% | 1.58% | 0.08% | 50.41% | 5.40% | 2.77% |
| Colemak Qi;x (2021) | 0.74% | 3.31% | 1.50% | 0.16% | 48.47% | 4.89% | 1.53% |
| Engram (2021) | 1.01% | 3.47% | 0.41% | 0.36% | 44.32% | 2.27% | 5.71% |
| ISRT (2021) | 0.65% | 3.50% | 1.51% | 0.28% | 50.70% | 5.16% | 2.41% |
| Semimak (2021) | 0.59% | 2.76% | 1.65% | 0.39% | 44.70% | 4.48% | 3.76% |
| APTv3 (2021) | 0.81% | 3.09% | 0.33% | 0.11% | 49.55% | 3.60% | 3.45% |
| Whorf (2021) | 0.45% | 2.73% | 1.66% | 0.77% | 46.67% | 4.42% | 3.74% |
| Nerps (2022) | 0.85% | 3.36% | 1.24% | 1.11% | 46.31% | 1.66% | 1.32% |
| CTGAP (2022) | 0.74% | 3.42% | 1.01% | 0.68% | 45.45% | 2.28% | 1.23% |
| Canary (2022) | 0.66% | 3.65% | 1.75% | 0.42% | 50.36% | 3.39% | 2.96% |
| Octa8 (2022) | 0.88% | 2.99% | 0.64% | 0.10% | 48.80% | 3.03% | 3.48% |
| Seht Drai (2022) | 0.63% | 2.74% | 0.83% | 0.64% | 57.84% | 5.59% | 3.75% |
| Sturdy (2022) | 0.62% | 2.80% | 1.58% | 0.42% | 50.10% | 2.85% | 2.09% |
| Gallium (2023) | 0.64% | 2.71% | 1.00% | 0.95% | 46.07% | 1.87% | 3.16% |
| Graphite (2023) | 0.68% | 2.73% | 0.87% | 0.41% | 46.01% | 1.80% | 2.34% |
| Recurva (2023) | 0.56% | 2.74% | 1.19% | 0.44% | 50.19% | 2.79% | 4.40% |
| Hanster-23 (2024) | 0.86% | 3.76% | 1.99% | 0.19% | 43.58% | 1.48% | 2.75% |
| Vylet (2024) | 0.92% | 3.08% | 0.75% | 0.07% | 49.23% | 3.73% | 3.44% |
| Focal (2024) | 0.53% | 3.23% | 0.99% | 0.46% | 47.00% | 2.93% | 2.09% |
Layouts that assign a letter to a thumb key:
| Layout | SFBs | SFSs | LSBs | Scissor | Rolls | Redir | Pinky off |
|---|---|---|---|---|---|---|---|
| Maltron (1977) | 0.66% | 2.74% | 0.70% | 0.11% | 50.89% | 6.62% | 4.86% |
| RSTHD (2016) | 0.70% | 2.94% | 0.81% | 0.09% | 53.20% | 6.40% | 0.79% |
| Vibranium (2023) | 0.57% | 3.20% | 0.36% | 0.44% | 46.93% | 1.29% | 4.11% |
| Promethium (2024) | 0.58% | 3.05% | 0.24% | 0.11% | 45.75% | 1.53% | 4.08% |
| Caster (2024) | 0.63% | 2.71% | 0.05% | 0.54% | 50.87% | 5.65% | 3.40% |
| Nordrassil (2024) | 0.77% | 3.48% | 0.85% | 0.33% | 43.63% | 1.10% | 1.88% |
| Night (2024) | 0.41% | 2.30% | 1.26% | 0.63% | 45.87% | 3.00% | 3.06% |
| Enthium v10 (2025) | 0.42% | 2.54% | 0.15% | 0.08% | 46.52% | 2.18% | 3.31% |
Notes:
These tables are periodically updated, computed by Cyanophage’s layout playground. Off-home pinky use is obtained by evaluating
(100*m_pinky_off/m_input_length)in the JS console.(2025-08-22) Metrics for all layouts were recomputed. The DSFBs column is now SFSs. The values for SFSs, Rolls, and Redirects have changed due to a fix in how these values are read from the analyzer and changes to the analyzer itself, but the overall layout rankings remain similar.
For each layout, the parenthesized year is my estimate when it was released. Take this with a grain of salt, information is often unfortunately limited.