diff --git a/ccn2019.rev3.Rmd b/ccn2019.rev3.Rmd
index 06a6d04..223277f 100644
--- a/ccn2019.rev3.Rmd
+++ b/ccn2019.rev3.Rmd
@@ -64,9 +64,11 @@
 #inspect_cor(NB,show_plot = T)
 ```
 
-```{r}
+```{r models}
 data <- seqs %>% 
   unnest(local_stats) %>%
+  # restructure correct column to avoid caret errors
+  # C stands for "CORRECT", and I is "INCORRECT"
   mutate(correct=factor(as.numeric(correct),labels=c("C","I"))) %>%
   filter(!is.na(correct), !is.na(rt))
 
@@ -74,12 +76,12 @@
 split <- nrow(data) * 0.8
 
 
-train <- data[1:split,]
-test <- data[(split+1):nrow(data),]
+train_data <- data[1:split,]
+test_data <- data[(split+1):nrow(data),]
 
-model <- train(
-  correct ~ .,
-  data = train,
+new_model <- train(
+  correct ~ .-rt-a-al,
+  data = train_data,
   method = "glm",
   family = "binomial",
   trControl = trainControl(
@@ -90,13 +92,36 @@
   )
 )
 
-model
+new_model
 
-p <- predict(model, test, type="prob")
+predicted_new_data <- predict(new_model, test_data, type="prob")
 
-confusionMatrix(ds$correct, prd)
+confusionMatrix(test_data$correct, predicted_new_data)
 
 library(caTools)
-colAUC(p,test$correct, plotROC=T)
-```
+colAUC(predicted_new_data, test_data$correct, plotROC=T)
 
+
+## OLD MODEL (only global features)
+old_model <- train(
+  correct ~ n+t+v,
+  data = train_data,
+  method = "glm",
+  family = "binomial",
+  trControl = trainControl(
+    method = "cv",
+    number = 5,
+    classProbs = T,
+    summaryFunction = twoClassSummary
+  )
+)
+
+old_model
+
+predicted_old_data <- predict(old_model, test_data, type="prob")
+
+confusionMatrix(test_data$correct, predicted_old_data)
+
+library(caTools)
+colAUC(predicted_old_data,test_data$correct, plotROC=T)
+```