jabowery
-Interested User-
Posts: 5
Joined: Sep 5, 2019
|
Posted: Nov 24, 2019 11:31 AM
Msg. 1 of 5
How can one detect when a symbol's price increase is due to a reverse split?
I've recently been bit by a backtest of an algorithmic trade with a ridiculously high yield that, when I tracked it down, was due to a reverse split causing an apparent increase in the stock's price.
|
DTN_Stephen
-DTN Guru-
Posts: 453
Joined: Aug 22, 2014
|
Posted: Nov 25, 2019 06:47 AM
Msg. 2 of 5
We send the last 2 split dates and rates through our Fundamental data fields.
Stephen Shockey Senior Customer Support Representative and Product Support Specialist
DTN 800-779-7299 stephen.shockey@dtn.com
|
jabowery
-Interested User-
Posts: 5
Joined: Sep 5, 2019
|
Posted: Nov 25, 2019 02:10 PM
Msg. 3 of 5
Is it then true, as seems to be the implication, that minute aggregated pricing cannot with certitude be relied upon for backtesting prior to the first of the 2 split dates provided in the Fundamental data fields? Edited by jabowery on Nov 25, 2019 at 02:18 PM
|
DTN_Gary_Stephen
-DTN Guru-
Posts: 396
Joined: Jul 3, 2019
|
Posted: Nov 26, 2019 01:49 PM
Msg. 4 of 5
Yes, that is a correct statement. If you like, I can put in a request to add more historical split dates to the data the API can return.
Sincerely, Gary Stephen DTN IQFeed Implementation Support Specialist
|
jabowery
-Interested User-
Posts: 5
Joined: Sep 5, 2019
|
Posted: Nov 26, 2019 04:21 PM
Msg. 5 of 5
Let me explain the use case so you can properly prioritize this data feature (which is available in its entirety at https://www.splithistory.com/): While backtesting an AlgoTerminal trading strategy, a market buy order is accompanied with a limit order. At the session close (end of day/EOD), the limit order is not reached and no bar update comes in at the end of day to permit the algorithm to sell the position at market. An overnight reverse X:1 split occurs but this is currently undetected by AlgoTerminal so it doesn't adjust the position (#shares/X and limit price*X). Upon resumption of trading the next day, the price is X*closing price > limit price. The limit condition met, AlgoTerminal sells #shares at price*X. Backtesting incorrectly reports a highly successful trading strategy. If the backtesting stream included split multiples as well as prices, AlgoTerminal could do the adjustments necessary to produce valid backtesting reports. Edited by jabowery on Nov 26, 2019 at 04:22 PM
|