diff --git a/cumulant_analysis.jl b/cumulant_analysis.jl
index 40cd4fd..b685404 100644
--- a/cumulant_analysis.jl
+++ b/cumulant_analysis.jl
@@ -2,6 +2,7 @@ cd("/share/tmschaef/jkott/modelB/KZ/cumulants")
 
 using Plots
 using DelimitedFiles
+using Statistics
 
 const L = 32
 const z = 3.906e0
@@ -19,28 +20,28 @@ const t_e = (m²e - m_b)/m_a
 const t_c_steps = trunc(Int, t_c/Δt)
 
 skip = 10
-maxt = (trunc(Int, t_e / Δt)+1)÷skip+1
+maxt = (trunc(Int, t_e / Δt)+1)÷skip+2
 
 id_max = 16
 series_max = 16
 run_max = 32
 n_files = id_max * series_max * run_max
 
-M4 = zeros(Float32, maxt, n_files)
-M2 = zeros(Float32, maxt, n_files)
+M4 = zeros(Float64, div(maxt,10)+1, n_files)
+M2 = zeros(Float64, div(maxt,10)+1, n_files)
 
 function cumulant_run(fn, i)
     df = readdlm(fn, ' ')
-    M = df[:,3]
+    M = df[1:10:end,3]
 
-    M4[:,i] .= M.^4
-    M2[:,i] .= M.^2
+    M4[:,i] .= (M*(L^3/2)^(1/4)).^4
+    M2[:,i] .= (M*(L^3/2)^(1/4)).^2
 end
 
 function collect_files()
     Threads.@threads for id in 1:id_max
         for series in 1:series_max, run in 1:run_max
-            cumulant_run("sum_L_$L"*"_id_$id"*"_series_$series"*"_run_$run.dat", run + (series-1)*series_max + (id-1)*series_max*id_max)
+            cumulant_run("sum_L_$L"*"_id_$id"*"_series_$series"*"_run_$run.dat", run + (series-1)*run_max + (id-1)*series_max*run_max)
         end
     end
 end
@@ -50,8 +51,8 @@ collect_files()
 M4_err = std(M4, dims=2) / sqrt(n_files)
 M2_err = std(M2, dims=2) / sqrt(n_files)
 
-M4 .= sum(M4, dims=2)/n_files
-M2 .= sum(M2, dims=2)/n_files
+M4 .= sum(M4, dims=2)
+M2 .= sum(M2, dims=2)
 
 c4 = M4 .- 3 * M2.^2
 c4_err = sqrt.(M4_err.^2 .+ (6 * M2 .* M2_err).^2)
@@ -59,10 +60,10 @@ c4_err = sqrt.(M4_err.^2 .+ (6 * M2 .* M2_err).^2)
 output_file = open("data_cumulant_$L.jl","w")
 
 write(output_file, "c4 = ")
-show(output_file, C_tot)
+show(output_file, c4)
 write(output_file, "\n\n")
 
 write(output_file, "c4_err = ")
 show(output_file, c4_err)
 
-close(output_file)
\ No newline at end of file
+close(output_file)