diff --git a/build.zig b/build.zig
index 36668b4..4617515 100644
--- a/build.zig
+++ b/build.zig
@@ -392,6 +392,10 @@ const exercises = [_]Exercise{
         .main_file = "076_sentinels.zig",
         .output = "Array:123056. Many-item pointer:123.",
     },
+    .{
+        .main_file = "076a_memory_allocation.zig",
+        .output = "Running Average: 0.30 0.25 0.20 0.18 0.22",
+    },
     .{
         .main_file = "077_sentinels2.zig",
         .output = "Weird Data!",
diff --git a/exercises/076a_memory_allocation.zig b/exercises/076a_memory_allocation.zig
new file mode 100644
index 0000000..b098773
--- /dev/null
+++ b/exercises/076a_memory_allocation.zig
@@ -0,0 +1,70 @@
+// In most of the examples so far, the inputs are known at compile time, thus
+// the amount of memory used by the program is fixed and is requested. However, if responding to
+// input whose size is not known at compile time, such as:
+//  - user input via command-line arguments
+//  - inputs from another program
+//
+// You'll need to request memory for you program to be allocated by your
+// operating system at runtime.
+//
+// Zig provides several different allocators. In the Zig documentation, it
+// recommends the Arena allocator for simple programs which allocate once and
+// then exit:
+//
+//     const std = @import("std");
+//
+//     // memory allocation can fail because your computer is out of memory, so
+//     // the return type is !void
+//     pub fn main() !void {
+//         
+//         var arena = std.heap.ArenaAllocator.init(std.heap.page_allocator);
+//         defer arena.deinit();
+//
+//         const allocator = arena.allocator();
+//
+//         const ptr = try allocator.create(i32);
+//         std.debug.print("ptr={*}\n", .{ptr});
+//
+//         const slice_ptr = try allocator.create(i32);
+//         std.debug.print("ptr={*}\n", .{ptr});
+//     }
+
+// Instead of a simple integer, this program requires a slice to be allocated that is the same size as an input array
+
+// Given a series of numbers, take the running average. In other words, the running average of the last N elements
+
+const std = @import("std");
+
+fn runningAverage(arr: []const f64, avg: [] f64) void {
+    var sum: f64 = 0;
+
+    for (0.., arr) |index, val| {
+        sum += val;
+        avg[index] = sum / @intToFloat(f64, index + 1);
+    }
+}
+
+pub fn main() !void {
+    // pretend this was defined by reading in user input
+    var arr: []const f64 = &[_]f64{ 0.3, 0.2, 0.1, 0.1, 0.4 };
+
+    // initialize the allocator
+    var arena = std.heap.ArenaAllocator.init(std.heap.page_allocator);
+
+    // free the memory on exit
+    defer arena.deinit();
+
+    // initialize the allocator (TODO: replace this with ???)
+    const allocator = arena.allocator();
+
+    // TODO: replace this whole line with ???
+    var avg = try allocator.alloc(f64, arr.len);
+
+    runningAverage(arr, avg);
+    std.debug.print("Running Average: ", .{});
+    for (avg) |val| {
+        std.debug.print("{d:.2} ", .{val});
+    }
+}
+
+// For more details on memory allocation and the different types of memory allocators, see https://www.youtube.com/watch?v=vHWiDx_l4V0