diff --git a/day16/day16.py b/day16/day16.py
index cfa1b10..bebb6e0 100644
--- a/day16/day16.py
+++ b/day16/day16.py
@@ -8,11 +8,12 @@ identify invalid nearby tickets by considering only whether tickets contain valu
 
 
 def main():
-    rules, my_ticket, tickets = open("input.txt").read().split("\n\n")
+    rules, my_ticket, other_tickets = open("input.txt").read().split("\n\n")
     rules = parse_fields(rules)
-    tickets = tickets.splitlines()[1:]
-    print("Ticket scanning error rate ", part1(tickets, rules))
-    part2(tickets, rules)
+    my_ticket = my_ticket.splitlines()[1]
+    other_tickets = other_tickets.splitlines()[1:]
+    print("Ticket scanning error rate ", part1(other_tickets, rules))
+    part2(my_ticket, other_tickets, rules)
 
 
 def parse_fields(fields):
@@ -35,31 +36,68 @@ def validate_ticket(ticket, rules):
     invalid_fields = []
     for value in ticket.split(","):
         value = int(value)
-        validations = (any(value in r for r in rule) for rule in rules.values())
-        if not any(validations):
+        if not validate_field(value, *rules.values()):
             invalid_fields.append(value)
     return invalid_fields
 
 
-def part2(tickets, rules):
-    # filter only valid tickets
-    valid_tickets = [ticket for ticket in tickets if validate_ticket(ticket, rules) == []]
+def validate_field(field, *rules):
+    validations = (any(field in r for r in rule) for rule in rules)
+    return any(validations)
 
-    # field => [matching_rules]
-    # eliminate field with only one matching rule
-    all_matching_rules = {}
-    for ticket in valid_tickets:
-        possible_matching_rules = defaultdict(set)
-        for field_index, field_value in enumerate(ticket.split(",")):
-            field_value = int(field_value)
-            for name, rule in rules.items():
-                match = {r: field_value in r for r in rule}
-                if any(match.values()):
-                    possible_matching_rules[name].add(field_index)
-        all_matching_rules[ticket]= possible_matching_rules
-    a = 1
+
+def part2(my_ticket, other_tickets, rules):
+    # filter only valid tickets
+    valid_tickets = [ticket for ticket in other_tickets if validate_ticket(ticket, rules) == []]
+    valid_tickets.append(my_ticket)  # my ticket is valid
+
+    # possible field for each index of a ticket
+    candidates = defaultdict(set)
+    for index in range(len(rules)):
+        def inner():
+            for rule_name, constraints in rules.items():
+                for ticket in valid_tickets:
+                    field_value = int(ticket.split(",")[index])
+                    if not validate_field(field_value, constraints):
+                        return
+                candidates[index].add(rule_name)
+        inner()
     
-    # repeat
+    sorted_candidates = sort_candidates(candidates)
+
+    fields_indexes = {}
+    try:
+        while len(fields_indexes) != len(rules):
+            index, found = sorted_candidates.popitem()
+            found = next(iter(found))
+            fields_indexes[index] = found
+            sorted_candidates = remove_item(sorted_candidates, found)
+    except:
+        pass
+
+    fields_indexes = {k: v for k,v in fields_indexes.items() if v.startswith('departure')}
+
+    total = 1
+    my_ticket = my_ticket.split(',')
+    for index in fields_indexes:
+        total *= int(my_ticket[index])
+    a = 1
+
+
+def sort_candidates(c):
+    return {x: c[x] for x in sorted(c, key=lambda k: len(c[k]), reverse=True)}
+
+def remove_item(candidates, item):
+    ret = {}
+    for key, value in candidates.items():
+        try:
+            value.remove(item)
+        except ValueError:
+            pass
+        ret[key] = value
+
+    #candidates = {k: set(v - item) for k,v in candidates.items()}
+    return ret