Revert "Bump verification width to 4"

[jackkleeman]

Reverts #146 because we're seeing some timeouts

[github-actions]

Test Results

  71 files  ±0    71 suites  ±0   7m 10s ⏱️ - 8m 41s
  58 tests ±0    57 ✔️ +1  1 💤 ±0  0 ❌  - 1 
151 runs  ±0  150 ✔️ +1  1 💤 ±0  0 ❌  - 1 

Results for commit dacca7a. ± Comparison against base commit fe2d609.

[tillrohrmann]

What is the cause of the timeouts @jackkleeman? I thought that we are capping the max sleep duration independent of the width and depth.

[jackkleeman]

Its not about the sleeps, just the fact that some command trees can be large - in principle the max command tree with 4 and 14 has 200 million commands (extremely unlikely). The commands also do not all happen concurrently, many lock the same keys. We can investigate further if needed
At some point it is probably worth rethinking the command tree generation to have more predictable size of the tree. But callculating duration will always be hard, imo