DYO #2 Fruit fly population investigation

Question:The question was how would the gender of the fruitflies affect the number of fruitflies after 3 weeks.
Hypothesis:So my hypothesis was that if i put 2 female fruitflies and 1 male fruitfly together then i will get 322 fruitflies in 21 days because the female fruitflies lay 20 eggs a day times 2 because the amount of females so you add that up and get 322.
Backround Research:when i researched i found out that female fruitflies store sperm from males and lay 20 eggs a day.i learned it takes 1 day for an egg to hatch, 7 days to become a pupa, 6 days to become a larva and 2 days to mature into an adult and be able to lay eggs. we also learned some flies cant fly because of genetics.
Procedure: 1st step. I chose 1 male and 2 female fruitflies.
2nd step. we prepared a vile with a food source.
3rd step.we put them in the vile sideways so they wouldnt fall into their food.
4th step. then we go and look at our vile and 2 other groups.
5th step.then in three we check to see how many flies are in each vile we chose.

this picture shows the graph and my data for this dy i picked because it was the only picture i had.

 

  Analysis: i looked at three vials that contained 1 male and 2 female fruit flies. each vial had different amount of fruit flies at the end of 3 weeks.vial 1 had 60 fruit flies. vile 2 had 105 fruit flies. vial 3 had 70 fruit flies.
if iput 1 male and 2 female fruit flies in avile with a food souce then in 21 days there will be 322 fruit flies. my data did not support my hypothesis because i messed up with my calculations but the average of the vials was 78.3. i did not calculate right because i didnt think of deaths and how eggs wouldnt hatch. i thought they would be born every day and thats how i got 322 but i was wrong.
Conclusion: the overall experiment was a complete fail thinking about my calculations there is no way to find out how it would happen with math there were to many unexpected deaths. i calculated 322 and only got 60 it was a fail my experiment didnt support the hypothesis. the best way to do this again is redo the experiment many time and average the deaths. the question i would like to go further with is how to find the best way to solve this problem.