docs: remove extra erroneous backtick

docs/index.md

@@ -4948,7 +4948,7 @@ When you spin up a process yourself you should generally have it pipe its output
 
 Go's built-in `testing` package provides support for running `Benchmark`s.  Earlier versions of Ginkgo subject-node variants that were able to mimic Go's `Benchmark` tests.  As of Ginkgo 2.0 these nodes are no longer available.  Instead, Ginkgo users can benchmark their code using Gomega's substantially more flexible `gmeasure` package.  If you're interested, check out the `gmeasure` [docs](https://onsi.github.io/gomega/#gmeasure-benchmarking-code).  Here we'll just provide a quick example to show how `gmeasure` integrates into Ginkgo's reporting infrastructure.
 
-`gmeasure` is structured around the metaphor of Experiments.  With `gmeasure` you create ``Experiments` that can record multiple named `Measurements`.  Each named `Measurement` can record multiple values (either `float64` or `duration`).  `Experiments` can then produce reports to show the statistical distribution of their `Measurements` and different `Measurements`, potentially from different `Experiments` can be ranked and compared.  `Experiments` can also be cached using an `ExperimentCache` - this can be helpful to avoid rerunning expensive experiments _and_ to save off "gold-master" experiments to compare against to identify potential regressions in performance - orchestrating all that is left to the user.
+`gmeasure` is structured around the metaphor of Experiments.  With `gmeasure` you create `Experiments` that can record multiple named `Measurements`.  Each named `Measurement` can record multiple values (either `float64` or `duration`).  `Experiments` can then produce reports to show the statistical distribution of their `Measurements` and different `Measurements`, potentially from different `Experiments` can be ranked and compared.  `Experiments` can also be cached using an `ExperimentCache` - this can be helpful to avoid rerunning expensive experiments _and_ to save off "gold-master" experiments to compare against to identify potential regressions in performance - orchestrating all that is left to the user.
 
 Here's an example where we profile how long it takes to repaginate books: