# Predictive formula

I would like to predict how much we should keep in a box "Magical-box"

"Magical-box" Should have the ability to predict the value of the next deposit in it :

For Example :

``````DepositNbr    Coins      MagicBox
1              6            6
2              4            4
3              10           13    <==> the prediction process may starts from the third deposit
4              13           8
5              8            23
6              23            2
7              2            ...
``````

is there any way to perform this prediction based on the past or the present ?,any formula ( markov Chain , normal distribution,Regression... ) is welcomed
###### Who is Participating?

x
I wear a lot of hats...

"The solutions and answers provided on Experts Exchange have been extremely helpful to me over the last few years. I wear a lot of hats - Developer, Database Administrator, Help Desk, etc., so I know a lot of things but not a lot about one thing. Experts Exchange gives me answers from people who do know a lot about one thing, in a easy to use platform." -Todd S.

Microsoft Excel ExpertCommented:
is this what you mean, plz see attached.
Book1.xlsb
Author Commented:
Not actually , I would like to get a predictive formula , mathematic one
"Batchelor", Developer and EE Topic AdvisorCommented:
Making predictions without knowing anything about the process is worthless. You have to adapt the model to the data ...
Author Commented:
the number of coins is generated randomly  with a normal distribution(average , stdev)
"Batchelor", Developer and EE Topic AdvisorCommented:
Then the best prediction is obviously to use the average of the cells above (all or in a set of 6 or 4 or whatever your horizon should be). For all cells above, the formula is   = Average(C\$2:C7). If you drag that down, C\$2 will stay, and C7 will change accordingly.
Commented:
>>  the number of coins is generated randomly  with a normal distribution(average , stdev)

Do you know the average and standard deviation for the process?
If so, you should always guess the average.
If not you should analyze the data stream to evaluate the average and then always use that as your prediction.

If the process is a PRNG implemented in software, you can try to crack it (look up or disassemble the code and identify the key).  If you were successful you would be able to make perfect predictions.

But what you are asking is very much like predicting the result of a series of coin flips.  You can't expect to be correct much more (or less) and 50% of the time.

Experts Exchange Solution brought to you by