Backtesting: Backtesting is re-running a new or changed pricing model against real past quotes to see how every price would have moved before the change goes live.
When you change a rate, a markup or a rule, you want to know what it does before buyers see it. Backtesting answers that question with your own history. Take a set of recent quotes, price each of them again with the draft model, and compare the old and new prices side by side.
Example: a shop plans to raise its SLS machine rate and lower its minimum charge. It backtests the draft against the last 200 quotes. The result shows that most prices move by a few percent, small orders get cheaper, and three large jobs rise by more than 15%. The shop decides to adjust the bulk discount before publishing.
Backtesting is not the same as calibration. Calibration changes the model so its estimates match what really happened on the machine. Backtesting does not change anything: it shows the effect of a change you are considering. A good routine uses both: calibrate the estimates, then backtest the new prices.
A backtest tells you how prices move, not how buyers would have reacted. Pair it with your win rate on each kind of job to judge whether a change is likely to help.
In Quotometry, configurations are versioned, and the simulator backtests a draft configuration against recent quotes before you publish it. See the simulator for how that works.
Go deeper: Quote win rate: why buyers go quiet after a quote, and what to measure