I'm running a MetaStrategy with 5 equally weighted, long-only child strategies. Each is set to use a custom dataset of 4 ETFs (the same four for each child strat). It runs without leverage so that I can have up to 5 positions for each ETF (one for each strat). It only trades once a day, at the open, based on previous EOD data.
When I ran it just now based on Friday's EOD, I get two SELL signals for strats #4 and #5 for SOXL. The trouble is, I only have one open position for SOXL, in strat #5. This is what I mean by a phantom sell signal. I had a position in SOXL from Jan 16 until Feb 4, so it's been closed a couple of days.
Now if I double-check by running strat #4 by itself (outside the MetaStrategy), it turns out that I still have a position in SOXL open... it opened a position on Jan 16 (just like the MetaStrategy, good), but the strat #4 did not close it on Feb 4. For some reason, the data is showing that it closed Feb 4 in the MetaStrategy, but didn't close in strat #4.
I'm not sure what to conclude. I think that the MetaStrategy *should* follow its own system, and not give me a phantom signal since it closed this trade on Feb 4. But instead for some reason, it's giving me the signal coming through from Strat #4 telling me to close my SOXL trade that I don't have open. Currently, I'm trading the signals manually, so I can ignore that sell signal. But what happens if I trade it automatically, and it sends a sell signal to my broker... my broker doesn't know these orders are coming from separate substrategies... it might see that I have available shares of SOXL and sell them off, even though they were there waiting for a different signal to come in at some later time.
Anybody else run into this? How do I investigate it further, or change some obscure setting to make sure it doesn't happen?
When I ran it just now based on Friday's EOD, I get two SELL signals for strats #4 and #5 for SOXL. The trouble is, I only have one open position for SOXL, in strat #5. This is what I mean by a phantom sell signal. I had a position in SOXL from Jan 16 until Feb 4, so it's been closed a couple of days.
Now if I double-check by running strat #4 by itself (outside the MetaStrategy), it turns out that I still have a position in SOXL open... it opened a position on Jan 16 (just like the MetaStrategy, good), but the strat #4 did not close it on Feb 4. For some reason, the data is showing that it closed Feb 4 in the MetaStrategy, but didn't close in strat #4.
I'm not sure what to conclude. I think that the MetaStrategy *should* follow its own system, and not give me a phantom signal since it closed this trade on Feb 4. But instead for some reason, it's giving me the signal coming through from Strat #4 telling me to close my SOXL trade that I don't have open. Currently, I'm trading the signals manually, so I can ignore that sell signal. But what happens if I trade it automatically, and it sends a sell signal to my broker... my broker doesn't know these orders are coming from separate substrategies... it might see that I have available shares of SOXL and sell them off, even though they were there waiting for a different signal to come in at some later time.
Anybody else run into this? How do I investigate it further, or change some obscure setting to make sure it doesn't happen?
Rename
Still unresolved, but a small update:
That strategy exits a trade when IBS(1) > 90. I considered the possibility that there was simply a rounding difference when calculating this comparison, where the MetaStrategy rounded one way, different from the child strategy. However, lowering the threshold to 89 and 88 caused no difference, and raising it to 91 also shows no difference... the MetaStrategy still shows that the previous trade closed on Feb 4, while the child strategy shows the trade continuing to be open.
So that's not the source of the open vs close problem, I don't think. But even if there WAS a small rounding error, I still don't like the whole idea of a MetaStrategy showing a sell signal for a position that isn't open in the MetaStrategy.
That strategy exits a trade when IBS(1) > 90. I considered the possibility that there was simply a rounding difference when calculating this comparison, where the MetaStrategy rounded one way, different from the child strategy. However, lowering the threshold to 89 and 88 caused no difference, and raising it to 91 also shows no difference... the MetaStrategy still shows that the previous trade closed on Feb 4, while the child strategy shows the trade continuing to be open.
So that's not the source of the open vs close problem, I don't think. But even if there WAS a small rounding error, I still don't like the whole idea of a MetaStrategy showing a sell signal for a position that isn't open in the MetaStrategy.
I didn’t think there are any kind of rounding errors kids you described, and i personally haven’t seen the kind of unsynchronized behavior you’re describing in my personal MetaStrategy that I’ve been trading for a few years.. Wothout seeing your extract setup and going through some investigation it’s not possible to say what might be happening. Would you be interested in a support call over Zoom? Write support@wealth-lab.com and we can set one up.
Will do.
Note to all: met with Glitch, shared strategies, replicated bug, build #158 out shortly should fix it.
Build 158 is out!
Your Response
Post
Edit Post
Login is required