The code below combines the other answers and adds a bit to number the vertices.

import arcpy
arcpy.env.workspace = "in_memory"
#paths
fc = r"...\polygons"
fc_out = r"...\vertices"
arcpy.MakeFeatureLayer_management(fc, "lyr")
# add fields if needed
for FIELD in ["DRAW_ORDER", "COUNT"]:
if FIELD not in [field.name for field in arcpy.ListFields(fc)]:
try:
arcpy.AddField_management("lyr", FIELD, "SHORT")
except Exception as e:
print e
# get the number of points minus overlapping (@dmahr - GSE)
arcpy.CalculateField_management("lyr", "COUNT", "!Shape!.pointCount-!Shape!.partCount", "PYTHON")
# dict to iterate and check count
OIDS = {}
for row in arcpy.da.SearchCursor("lyr", ["OBJECTID", "COUNT"]):
OIDS[row[0]] = row[1]
del row
# get vertices as points and add XY (@Aaron - GSE)
arcpy.FeatureVerticesToPoints_management("lyr", fc_out)
arcpy.AddXY_management(fc_out)
# start adding a number to the points
for OID in OIDS:
order_count = 1
rows = arcpy.da.UpdateCursor(fc_out, ["DRAW_ORDER", "COUNT"], "ORIG_FID = %d"%OID)
for row in rows:
# will leave the overlapping as NULL
if order_count <= OIDS[OID]:
row[0] = order_count
rows.updateRow(row)
order_count += 1
## # this can set the overlapping to 0 or some unique value (999)
## else:
## row[0] = 0
## rows.updateRow(row)
The points are labelled in drawing order. The last point (under the first) will have no label and can be deleted by selecting all points that have Null, or unique, "DRAW_ORDER" values if not needed for reconstruction. A definition query can be used to remove the overlapping points from display.
XY data is present, but I will leave that to your labeling/displaying desires. See Aaron's answer about adding a XY field for labeling.
I was also toying with FeatureClass to numpy array, but I finished this first.