Test adding figure

_episodes/03-making_useful_figures.md

@@ -24,24 +24,28 @@ The plots created in a benchmark should be designed to be legible not only to yo
 
 Ideally, benchmark designers should aim to design paper-ready figures labeled with large clear text and a [perceptual color palette](https://root.cern/blog/rainbow-color-map/). Remember that we want to create figures that can be used as-is in the TDR, in conference presentations, and in evaluating detector and software performance.
 
+An example figure from the u-channel rho benchmark is shown here:
+
+![Example benchmark figure]({{ page.root }}/fig/example_benchfig.png)
+
 In this example, the plot is labeled with the collision system and other details:
 ```c++
-TLatex* r42 = new TLatex(0.18, 0.91, "ep 10#times100 GeV");
-TLatex* r43 = new TLatex(0.9,0.91, "#bf{EPIC}");
-TLatex* r44 = new TLatex(0.53, 0.78, "10^{-3}<Q^{2}<10 GeV^{2}, W>2 GeV");
-TLatex* r44_2 = new TLatex(0.5, 0.83, ""+vm_label+" #rightarrow "+daug_label+" eSTARlight");
+TLatex* beamLabel       = new TLatex(0.18, 0.91, "ep 10#times100 GeV");
+TLatex* epicLabel       = new TLatex(0.9,0.91, "#bf{EPIC}");
+TLatex* kinematicsLabel = new TLatex(0.53, 0.78, "10^{-3}<Q^{2}<10 GeV^{2}, W>2 GeV");
+TLatex* generatorLabel  = new TLatex(0.5, 0.83, ""+vm_label+" #rightarrow "+daug_label+" eSTARlight");
 ```
 The axis labels and titles were scaled for clarity:
 ```c++
-base1->GetYaxis()->SetTitleSize(base2->GetYaxis()->GetTitleSize()*1.5);
-base1->GetYaxis()->SetLabelSize(base2->GetYaxis()->GetLabelSize()*1.5);
+histogram->GetYaxis()->SetTitleSize(histogram->GetYaxis()->GetTitleSize()*1.5);
+histogram->GetYaxis()->SetLabelSize(histogram->GetYaxis()->GetLabelSize()*1.5);
 ```
 The number of axis divisions was decreased to reduce visual clutter:
 ```c++
-h_effEtaPtPi->GetXaxis()->SetNdivisions(5);
+histogram->GetXaxis()->SetNdivisions(5);
 ```
 And margins were increased to allow for more space for larger axis labels:
 ```c++
-virtualPad1->SetBottomMargin(0.2);
+virtualPad->SetBottomMargin(0.2);
 ```