diff --git a/README.md b/README.md
index 38a51ef..320d727 100644
--- a/README.md
+++ b/README.md
@@ -1,2 +1,40 @@
 # CC-CH-pip-ana
-Analysis code for single charged pion production by muon neutrinos in the MINERvA CH detector.
+Analysis code for single charged pion production by muon neutrinos in the MINERvA CH detector and in the "medium energy" NuMI beam era.
+
+The main branch now uses MasterAnaDev (MAD) tuples as the input, general-reconstruction event record. And it also makes nearly complete use of the MINERvA Analysis Toolkit (MAT) to perform the full cross section calculation with systematics.
+
+## Reproducing Cross Section Results:
+1. **Create MC histograms** -- Loop over MC tuple files to create all needed truth and reco-MC histograms (~1-3 hours).
+2. **Create data histograms** -- Loop over data tuple files to create all needed data histograms (~30 minutes).
+3. **Combine MC and data histograms into a cross section** --  background, unfolding, efficiency and normalization (~ seconds).
+4. **Plot results** -- (~seconds).
+
+Step 1 involves running the script `xsec/makeCrossSectionMCInputs.C`. This script could be run over all playlists on a single machine over the course of days. Instead, `ProcessCCPiMacro.py` parallelizes the task by submitting to the Fermilab grid `makeCrossSectionMCInputs` jobs over single input tuples. This way, all jobs start and finish in a few hours, and they only need to be merged after using `MAT/macros/madd.cxx` (which is like ROOTs `hadd` except it merges root files containing `MnvHND` histograms.
+
+Steps 2 and 3 are combined into a single script: `crossSectionDataFromFile.C`. This script first performs step 2, which takes less than an hour on a single machine, and then does step 3 immediate after. Input is the MC input file resulting from step 1.
+
+Step 4 takes the input file from step 3 and just takes a few seconds to run on a single machine. All cross section steps -- including event selection, sideband tune, background subtraction, migration matrix, unfolding, efficiency, and cross sections -- are plotted along with systematics in many variables.
+
+## Reproducing Cross Section Diagnostic Results:
+Other plots, not strictly apart of the cross section calculation pipeline, are made by running other scripts, mostly in the studies `studies/` folder. They include cut-by-cut event selection truth breakdowns (`runCutVars`), efficiency-purity counts table (`runEffPurTable`), sideband breakdowns (`runSidebands`), and background truth breakdowns (`runBackgrounds`). There is also `xsec/binningStudy`. The closure test is a three-script process (1. modified `xsec/makeCrossSectionMCInputs.C`, 2. `runXSecLooper.cpp` and 3. `xsec/GXSEClosure.C`) that will be described in a separate section. The warping study to test unfolding just involves running the cross section pipeline with modifications to `xsec/makeCrossSectionMCInputs.C`.
+
+## Setup
+
+* Install [MAT-MINERvA](https://github.com/MinervaExpt/MAT-MINERvA) and [UnfoldUtils](https://github.com/MinervaExpt/UnfoldUtils).
+  * [GENIEXSecExtract](https://github.com/MinervaExpt/GENIEXSecExtract) is only needed if you will be performing closure tests.
+  * [MAT](https://github.com/MinervaExpt/MAT) is automatically installed by `MAT-MINERvA`
+```
+> tree your_working_dir/ -L 1
+├── CC-CH-pip-ana
+├── GENIEXSecExtract
+├── MAT
+├── MAT-MINERvA
+├── opt
+└── UnfoldUtils
+```
+* Set your `$TOPDIR` environment variable `export TOPDIR=/path/to/your_working_dir/`
+* Run other environmental setup: `source permissions.sh; source setSL7.sh`
+
+## Notes
+* lots more details to be added here.
+* legacy instructions are [here](https://docs.google.com/document/d/1fC22eg0K82p5Jc_UwUKmOgB-SSq5zDVDBe0p1iXLvxQ/edit?usp=sharing). They still give accurate direction for running scripts, but they're setup section is out of date.