Two PostgreSQL tips that circulate in 1C circles, raising shared_buffers and switching off synchronous_commit, got a proper workout. The benchmark used was the Gilev test, a common yardstick for measuring 1C performance. The run-to-run variation turned out to be the real story, not the parameters. Five repetitions on an unchanged configuration spread wider than the gap between any of the tested settings. In plain terms, the effect credited to a tweak can hide inside ordinary measurement noise. The article lays out the result tables and includes the commands needed to change each parameter. There is also a six-step procedure for readers who want to run their own measurement. One caveat deserves mention: disabling synchronous_commit trades durability for speed, and the takeaway is to repeat tests often enough to separate a real gain from scatter.