diff --git a/gtsam/linear/HessianFactor.cpp b/gtsam/linear/HessianFactor.cpp
index 734c3fb72..9c57ea5ea 100644
--- a/gtsam/linear/HessianFactor.cpp
+++ b/gtsam/linear/HessianFactor.cpp
@@ -210,12 +210,12 @@ namespace gtsam {
 
   /* ************************************************************************* */
   double HessianFactor::error(const VectorValues& c) const {
-    return ublas::inner_prod(c.vector(),
+    return 0.5 * (ublas::inner_prod(c.vector(),
         ublas::prod(
             ublas::symmetric_adaptor<const constBlock,ublas::upper>(info_.range(0, this->size(), 0, this->size())),
             c.vector())) -
         2.0*ublas::inner_prod(c.vector(), info_.rangeColumn(0, this->size(), this->size(), 0)) +
-        info_(this->size(), this->size())(0,0);
+        info_(this->size(), this->size())(0,0));
   }
 
 void HessianFactor::updateATA(const HessianFactor& update, const Scatter& scatter) {
diff --git a/gtsam/linear/HessianFactor.h b/gtsam/linear/HessianFactor.h
index 2906a231a..c1ca9616d 100644
--- a/gtsam/linear/HessianFactor.h
+++ b/gtsam/linear/HessianFactor.h
@@ -47,7 +47,7 @@ namespace gtsam {
     typedef MatrixColMajor InfoMatrix;
     typedef SymmetricBlockView<InfoMatrix> BlockInfo;
 
-    InfoMatrix matrix_; // The full information matrix, s.t. the quadratic error is [x -1]'*H*[x -1]
+    InfoMatrix matrix_; // The full information matrix, s.t. the quadratic error is 0.5*[x -1]'*H*[x -1]
     BlockInfo info_;    // The block view of the full information matrix.
 
     void updateATA(const JacobianFactor& update, const Scatter& scatter);
@@ -68,14 +68,14 @@ namespace gtsam {
     /** Construct a unary factor.  G is the quadratic term (Hessian matrix), g
      * the linear term (a vector), and f the constant term.  The quadratic
      * error is:
-     * f - 2*x'*g + x'*G*x
+     * 0.5*(f - 2*x'*g + x'*G*x)
      */
     HessianFactor(Index j, const Matrix& G, const Vector& g, double f);
 
     /** Construct a binary factor.  Gxx are the upper-triangle blocks of the
      * quadratic term (the Hessian matrix), gx the pieces of the linear vector
      * term, and f the constant term.  The quadratic error is:
-     * f - 2*x1'*g1 - 2*x2'*g2 + x1'*G11*x1 + x1'*G12*x2
+     * 0.5*(f - 2*x1'*g1 - 2*x2'*g2 + x1'*G11*x1 + x1'*G12*x2)
      */
     HessianFactor(Index j1, Index j2,
         const Matrix& G11, const Matrix& G12, const Vector& g1,
@@ -98,7 +98,7 @@ namespace gtsam {
     virtual void print(const std::string& s = "") const;
     virtual bool equals(const GaussianFactor& lf, double tol = 1e-9) const;
 
-    virtual double error(const VectorValues& c) const; /** [x -1]'*H*[x -1] (also see constructor documentation) */
+    virtual double error(const VectorValues& c) const; /** 0.5*[x -1]'*H*[x -1] (also see constructor documentation) */
 
     /** Return the dimension of the variable pointed to by the given key iterator
      * todo: Remove this in favor of keeping track of dimensions with variables?
diff --git a/gtsam/linear/tests/testHessianFactor.cpp b/gtsam/linear/tests/testHessianFactor.cpp
index 64ae7b845..54736b974 100644
--- a/gtsam/linear/tests/testHessianFactor.cpp
+++ b/gtsam/linear/tests/testHessianFactor.cpp
@@ -31,7 +31,6 @@ using namespace gtsam;
 using namespace std;
 
 /* ************************************************************************* */
-#ifdef BROKEN // because accesses keys_, now private
 TEST(HessianFactor, ConversionConstructor) {
 
   HessianFactor expected;
@@ -84,9 +83,13 @@ TEST(HessianFactor, ConversionConstructor) {
 
   HessianFactor actual(combined);
 
+  VectorValues values(std::vector<size_t>(dims, dims+2));
+  values[0] = Vector_(2, 1.0, 2.0);
+  values[1] = Vector_(4, 3.0, 4.0, 5.0, 6.0);
+
+  DOUBLES_EQUAL(combined.error(values), actual.error(values), 1e-9);
   EXPECT(assert_equal(expected, actual, 1e-9));
 }
-#endif
 
 /* ************************************************************************* */
 TEST(HessianFactor, Constructor1)
@@ -105,7 +108,7 @@ TEST(HessianFactor, Constructor1)
 
   HessianFactor factor(0, G, g, f);
 
-  double expected = 160.75;
+  double expected = 80.375;
   double actual = factor.error(dx);
 
   DOUBLES_EQUAL(expected, actual, 1e-10);
@@ -134,7 +137,7 @@ TEST(HessianFactor, Constructor2)
 
   HessianFactor factor(0, 1, G11, G12, g1, G22, g2, f);
 
-  double expected = 181.0;
+  double expected = 90.5;
   double actual = factor.error(dx);
 
   DOUBLES_EQUAL(expected, actual, 1e-10);