Beta Testing Siamond

Siamond
Posts: 278
Joined: Wed Jul 31, 2013 9:24 pm

Re: Beta Testing Siamond

Postby Siamond » Tue Aug 11, 2015 3:36 pm

Hm, I thought the Open Source project moved to a Jan 1st kind of math?

BUG: using the inflation-adjusted simplest simulation, I would have expected the first data point of any cycle to be (portfolio=$1M minus $40k; spending = $40k). And yet, this is not what happens...

Ok, I'll stop there - can't go any further without logs.

User avatar
bo_knows
Site Admin
Posts: 660
Joined: Fri Jun 28, 2013 4:33 pm

Re: Beta Testing Siamond

Postby bo_knows » Tue Aug 11, 2015 4:25 pm

Siamond wrote:Well, I wanted to start checking a few logs, but... I really need the 'specific years' feature to work to compare to my own Excel spreadhseet... :geek:

I tried the single cycle, it works, but... ahem, I found another issue (same issue with a regular all-years simulation):

BUG: nothing happens with the Download CSV box! I tried with both Chrome and Safari.

REQUEST: move the "Download CSV" box right by the "Close" box.


Yeah, I was trying to mess with that last night. The code that I use for making the CSV file was lifted from someone else. For whatever reason, if I change the "Download CSV" from a link to a button (which is more desirable for looks) it ceases to function.

Short story: I'm working on it :)
-Bo (Creator and Admin for cFIREsim)

User avatar
bo_knows
Site Admin
Posts: 660
Joined: Fri Jun 28, 2013 4:33 pm

Re: Beta Testing Siamond

Postby bo_knows » Tue Aug 11, 2015 6:21 pm

Siamond wrote:Hm, I thought the Open Source project moved to a Jan 1st kind of math?

BUG: using the inflation-adjusted simplest simulation, I would have expected the first data point of any cycle to be (portfolio=$1M minus $40k; spending = $40k). And yet, this is not what happens...

Ok, I'll stop there - can't go any further without logs.


First off, I believe it does represent the "Jan 1st kind of math". Secondly, a workaround to view more detailed stats is to hit Ctrl+Shift+I and look in the javascript console after a simulation. The "Results:" array provides each simulation iteration, and each year of the iteration.

You might be right about the first data point. Currently, for a given year: we take the portfolio start value, subtract out spending, and add in portfolio adjustments before doing market gains. Really, that value should actually BE the portfolio start value, eh?
-Bo (Creator and Admin for cFIREsim)

User avatar
bo_knows
Site Admin
Posts: 660
Joined: Fri Jun 28, 2013 4:33 pm

Re: Beta Testing Siamond

Postby bo_knows » Tue Aug 11, 2015 6:43 pm

By the way, I added many of your bugs/requests to the "Issues" tab on the GitHub repo for tracking.
-Bo (Creator and Admin for cFIREsim)

Siamond
Posts: 278
Joined: Wed Jul 31, 2013 9:24 pm

Re: Beta Testing Siamond

Postby Siamond » Tue Aug 11, 2015 8:48 pm

bo_knows wrote:
Siamond wrote:Hm, I thought the Open Source project moved to a Jan 1st kind of math?

BUG: using the inflation-adjusted simplest simulation, I would have expected the first data point of any cycle to be (portfolio=$1M minus $40k; spending = $40k). And yet, this is not what happens...

Ok, I'll stop there - can't go any further without logs.


First off, I believe it does represent the "Jan 1st kind of math". Secondly, a workaround to view more detailed stats is to hit Ctrl+Shift+I and look in the javascript console after a simulation. The "Results:" array provides each simulation iteration, and each year of the iteration.

You might be right about the first data point. Currently, for a given year: we take the portfolio start value, subtract out spending, and add in portfolio adjustments before doing market gains. Really, that value should actually BE the portfolio start value, eh?

Actually, I misspoke, and I agree with your last point. The portfolio value on Jan 1st of the first year should really be the starting value, no more, no less (therefore $1M with default settings). The spend for the year is really the *budgeted* spend for the year (therefore $40k with default settings), but it is not withdrawn yet (I tend to think of it as: we do the math on Jan 1st, and budgeted money is withdrawn and moves to a checking/savings account on Jan 2nd). On Jan 1st of the 2nd year, the portfolio value becomes (starting-value - spend-previous-year) * (1+returns), and the budgeted spend should be computed on this basis (for the algorithms depending on the current portfolio basis), and on the basis of the inflation of the previous year (for the algorithms requiring it). Etc. Make sense?

User avatar
bo_knows
Site Admin
Posts: 660
Joined: Fri Jun 28, 2013 4:33 pm

Re: Beta Testing Siamond

Postby bo_knows » Wed Aug 12, 2015 1:47 pm

Siamond wrote:The spend for the year is really the *budgeted* spend for the year (therefore $40k with default settings), but it is not withdrawn yet (I tend to think of it as: we do the math on Jan 1st, and budgeted money is withdrawn and moves to a checking/savings account on Jan 2nd). On Jan 1st of the 2nd year, the portfolio value becomes (starting-value - spend-previous-year) * (1+returns), and the budgeted spend should be computed on this basis (for the algorithms depending on the current portfolio basis), and on the basis of the inflation of the previous year (for the algorithms requiring it). Etc. Make sense?


I don't think that the way you described it is any different from what it does now.

Everything regarding spending is calculated based on the portfolio.start value of a given year, which is always equal to the portfolio.end value of the previous year.
-Bo (Creator and Admin for cFIREsim)

Siamond
Posts: 278
Joined: Wed Jul 31, 2013 9:24 pm

Re: Beta Testing Siamond

Postby Siamond » Wed Aug 12, 2015 11:51 pm

bo_knows wrote:
Siamond wrote:The spend for the year is really the *budgeted* spend for the year (therefore $40k with default settings), but it is not withdrawn yet (I tend to think of it as: we do the math on Jan 1st, and budgeted money is withdrawn and moves to a checking/savings account on Jan 2nd). On Jan 1st of the 2nd year, the portfolio value becomes (starting-value - spend-previous-year) * (1+returns), and the budgeted spend should be computed on this basis (for the algorithms depending on the current portfolio basis), and on the basis of the inflation of the previous year (for the algorithms requiring it). Etc. Make sense?


I don't think that the way you described it is any different from what it does now.

Everything regarding spending is calculated based on the portfolio.start value of a given year, which is always equal to the portfolio.end value of the previous year.

Well, I can't reproduce it now because the test version today seems a little bit broken (the Run Simulation box doesn't do anything!), but the portfolio value on the first year with default parameters was NOT $1M, so something is (was?) amiss.

Siamond
Posts: 278
Joined: Wed Jul 31, 2013 9:24 pm

Re: Beta Testing Siamond

Postby Siamond » Sun Aug 23, 2015 10:16 am

Siamond wrote:
bo_knows wrote:
Siamond wrote:The spend for the year is really the *budgeted* spend for the year (therefore $40k with default settings), but it is not withdrawn yet (I tend to think of it as: we do the math on Jan 1st, and budgeted money is withdrawn and moves to a checking/savings account on Jan 2nd). On Jan 1st of the 2nd year, the portfolio value becomes (starting-value - spend-previous-year) * (1+returns), and the budgeted spend should be computed on this basis (for the algorithms depending on the current portfolio basis), and on the basis of the inflation of the previous year (for the algorithms requiring it). Etc. Make sense?


I don't think that the way you described it is any different from what it does now.

Everything regarding spending is calculated based on the portfolio.start value of a given year, which is always equal to the portfolio.end value of the previous year.

Well, I can't reproduce it now because the test version today seems a little bit broken (the Run Simulation box doesn't do anything!), but the portfolio value on the first year with default parameters was NOT $1M, so something is (was?) amiss.

I tried again, and using default parameters, the same thing occurs. I have a bit of a suspicion that the portfolio value by the end of the year is displayed, instead of the portfolio value at the beginning of the year - not sure though.

User avatar
bo_knows
Site Admin
Posts: 660
Joined: Fri Jun 28, 2013 4:33 pm

Re: Beta Testing Siamond

Postby bo_knows » Mon Aug 24, 2015 1:20 am

Siamond wrote:I tried again, and using default parameters, the same thing occurs. I have a bit of a suspicion that the portfolio value by the end of the year is displayed, instead of the portfolio value at the beginning of the year - not sure though.


The graphs ARE the end of year portfolio numbers, do you think it should show beginning of year? why?
-Bo (Creator and Admin for cFIREsim)

User avatar
bo_knows
Site Admin
Posts: 660
Joined: Fri Jun 28, 2013 4:33 pm

Re: Beta Testing Siamond

Postby bo_knows » Mon Aug 24, 2015 12:47 pm

Also, I'm not sure what URL you've been trying, but I just updated the version at http://104.131.112.188/input.php with work I've done. Output graphs show up in tabs. Stats are there. Still trying to work through a bug in the output graphs. Things are getting closer.
-Bo (Creator and Admin for cFIREsim)


Return to “cFIREsim Open Source Project”

Who is online

Users browsing this forum: No registered users and 8 guests