I started automation with Selenium WebDriver and later moved most new work to Playwright. Both are good. They're good at different things.
Where Selenium Still Wins
- Real browser/grid diversity: if you need to test on a specific old Safari or a vendor cloud grid, Selenium's ecosystem is unmatched.
- Language spread: mature bindings for Java, Python, C#, Ruby, JS — useful when your team isn't JS-first.
- Existing investment: a large, stable Selenium suite is not worth rewriting just for fashion.
Where Playwright Wins
- Auto-waiting: it waits for elements to be actionable before interacting, which kills a whole class of flaky
sleep()hacks. - Speed: it drives the browser over the DevTools protocol, not the slower WebDriver wire protocol.
- Built-in superpowers: network interception, tracing, video, and parallel isolation come in the box.
The Flakiness Difference
This is the practical one. In Selenium I wrote a lot of explicit waits:
new WebDriverWait(driver, Duration.ofSeconds(10))
.until(ExpectedConditions.elementToBeClickable(button));
button.click();
In Playwright the wait is implicit:
await page.getByRole('button', { name: 'Submit' }).click();
The locator retries until the element is actionable or it times out. Fewer waits means fewer flaky failures, which means people actually trust the suite.
My Default Now
For a new project with a JS/TS team, I start with Playwright — the flakiness savings alone justify it. For a team standardized on Java with a working grid, I stay on Selenium. The framework that your team will maintain beats the framework that benchmarks best.
The Part Neither Solves
Both let you write brittle, over-coupled tests. Locator strategy, test isolation, and not asserting on volatile UI text matter more than the tool. A disciplined Selenium suite beats a careless Playwright one every time.