Incomplete Statements... Live and Simulation

In live Portfolio, some stock selections have Incomplete Statements.
These stocks can be declined or accepted into portfolio.

In Simulation Backtesting, are Incomplete Statements purchased into the backtest?

If so, can they be eliminated from selection, and if so… How ?

Also, if a statement is updated ( completed ) , can this lead to a change of stocks included
in a simulation backtest, different stocks added / some stocks deleted… due to a change
of statement. test results different with updated statements ???

thanks,

At any historic point in time, any new data from incomplete statements is included in the system data. Any statement line item that is NA will retain the older existing values. To minimize the number of NAs in Sims P123 falls back to the latest prior statement with a good value (there are a few exceptions). So the Sim will buy all stocks that are at that time highest ranked and meet all the buy rules. There is no way to reject stocks with incomplete statements in Sims.

If a stock has an updated statement the new data values will be used in calculating the stock’s rank at the next rebalance. If the new rank is lower, and you have a rank sell rule with a value above the stock’s new rank value, it will be sold. if you have other sell rules that tests line items and a new value from the revised statement meets the sell rule criteria, the stock will be sold. In other words, a stock will be sold only if a sell rule is triggered by new values from the revised statement.

I think the above reflects what I as an investor would do if analyzing data by hand without the insight provided by P123. When new data is available from incomplete statements, I would compare the values to the older values I have for the stock and evaluate the differences to see if any of them justify a buy or sell of the stock. I would never evaluate a stock only based on the data in an incomplete statement.

Denny :sunglasses:

so if my Live Portfolio system selects 10 stocks,
and 2 or 3 selections have Incomplete Statements,
and are then excluded…replaced with stocks having
complete statements…

how can my Backtest Simulation…with new and improved statements…
that i am now trading live…ever be conforming to what takes place in real time ?

there is no way to test without having the advantage of the new completed statement.

will not the backtest almost always have an unknown advantage in performance,
to then underperform in Live ?

Jim, you can test it. Add a rule in your simulation

CompleteStmt = TRUE

" Jim, you can test it. Add a rule in your simulation " CompleteStmt=TRUE

is not the issue… when you run a simulation… The Completed Statement will already be contained for stocks selected…
and that correction of data is in hindsight,
thus… simulations have the correct data, Trading Live does not…thus your expectations and current trading
will probably never match the corrected data used in simulations?

this is not a flaw with P 1.2.3., nor is it meant to deceive.

is not the data used in simulations “corrected” / “updated” as it unintentionally selects stocks that
might be a better candidate for owning AFTER the updated data / statements are known ?

is it possible to run simulations that do not allow for the updating of an incomplete statement ?

if backtesting apples, i do not want to trade oranges… hoping to get apples.

did any of this make sense, or is my senility increasing ?

community help / explanation is always appreciated…

Chipper6,

please clarify CompleteStmt = TRUE

if I backtest in Sim adding this rule, and test just the year 2002,
will stocks qualify that only had Complete Statements during 2002,
or will hindsight update any incomplete statement, to now make it complete
for backtesting of the year 2002, and so on?

I am looking for stocks that never had incomplete statements at any point in history,
to avoid inaccurate test results…due to changing information.

does this avoid / eliminate prior incomplete statements from turning into completed statements?

thanks…

It’s supposed to be point in time. That means that at any given point in the past the backtest is only supposed to “see” what it would have seen back then if you would have had a time machine. Therefore if on a given day there was no complete statement then it won’t turn into one retroactively.

thank you for the clarification.

Jim,

You responded after my post above; “so if my Live Portfolio system selects 10 stocks, and 2 or 3 selections have Incomplete Statements, and are then excluded…replaced with stocks having complete statements…”

You may not have understood my post. in your live Port, the stocks that have incomplete statements for certain line items, would have fallback values for all the NA values in the incomplete statement. Instead of the NAs, it would use the values from the previous complete statement. The only remaining NA would be those line items that were also NA in the latest complete statement. Those would likely have been NA also when the stock was originally bought. Some types of industries report no values for certain line items. So for those industries, those line items are always NA.

So the Port would have complete and latest rank values for the stocks, and would not recommend a sell due to NA values in an incomplete statement that wasn’t also NA in the prior complete statement. In this respect, Sims and Ports handles NAs the same way so you don’t need to try to simulate a different way the Port handles NAs.

Denny :sunglasses: