If an account rises to 10,000, falls to 8,500 and then recovers, the drawdown at the low was 1,500, or 15% of the peak. Measured in R, the same idea tracks the fall in cumulative R from the highest point to the current one.
Drawdowns are not a sign that something is broken. Any method with a positive expectancy will have them, and their depth depends on variance, win rate and position size. What matters is whether the depth you actually experience is within what your plan expected.
They matter because they test behaviour. A deep drawdown is when many traders abandon a sound method, double size to recover, or begin to revenge trade. Knowing in advance how large a drawdown your numbers could produce lets you decide, while calm, what you will do at each depth.
Recovery is also asymmetric. A 10% loss needs an 11.1% gain to return to the peak, and a 50% loss needs a 100% gain. That arithmetic is the main argument for limiting risk per trade.
Many plans include a drawdown rule: at a set depth, cut size or stop trading and review. It converts a stressful decision into a prepared one.