StockFetcher Forums · Filter Exchange · BIGG Brexit Meeting next week...how is everyone playing it?<< 1 ... 3 4 5 6 7 >>Post Follow-up
StefanM
25 posts
msg #130151
Ignore StefanM
7/29/2016 8:50:56 AM

Yes of course.. thank you guys .... :-) !

c1916
70 posts
msg #130550
Ignore c1916
8/24/2016 4:14:59 PM

@Kevin_in_GA I've been following the VXZ/VIXZ pairs for the past couple of months. I've paper traded since we recently had this conversation, and I'm even further away from clarity on the subject than back in June (although I am completely comfortable with the logic at this point). My guess is that PTL (which I've been using) doesn't have the details that PTF (which you reference) uses, therefore I'm not seeing as many entry points as your June example suggests "should" come into play. For example, I only hit two round trips TOTAL over the past two months, and I'm running those trades off a 1.0 delta (less restrictive than your 2.5 delta), which suggests that I should have seen MORE activity. Additionally, I haven't seen a single entry point for more than 2 weeks, which just doesn't seem right.

Have you continued down this path since June?
Are you willing to discuss it further on/off line?

ferndave
54 posts
msg #130551
Ignore ferndave
8/24/2016 5:03:29 PM

c1916: I haven't traded that specific pair, but I have paper traded other VIX pairs for a few months. PTL's backtests use end of day numbers and assumes you filled every order perfectly that second. Live trading differs. I won't speak for Kevin, but my paper trades using VIX combinations have been less than promising. They are often too closely correlated resulting in quick round trips, usually for a loss. I ditched them because they weren't worth it.

Kevin_in_GA
4,548 posts
msg #130552
Ignore Kevin_in_GA
8/24/2016 5:22:58 PM

PTL does one thing right the PairTradeFinder does not - it calculates the Z-score based on the current bid/ask spread, not on the close. That should remove spread costs from the equation, but what I found (same as Dave) was that the SD size was too small, and one penny move in either direction could trigger an exit, usually at a loss. This was a real learning experience for me - basically that pair trades must have a sufficiently large SD as a percentage of the MA or you cannot really make decent cash. The result is that there are no intraday trade opportunities that meet that requirement (even if PairTradeFinder was "trading" them).

You need to be looking for pairs that are not overly correlated, and where the 2SD position is at least a few percent above the MA. These trades take days to weeks to close, and you must pay for 4 trades versus the traditional 2 trades for a directional play.

c1916
70 posts
msg #130562
Ignore c1916
8/24/2016 10:52:17 PM

Thanks for the comments.

@ferndave, by tying the PTL portfolio to an IB account, the theory is that one is indeed filling the order at the precise second that the trigger is reached. The issue I've found is that the trigger doesn't seem to be triggering as expected. Clearly I have something incorrect in my portfolio set up. I do not believe that timing of execution is the issue here, getting to the point of actual executing is my issue.

My goal here was to generate PTL code (bound to an IB account) that will result in the current-term equivalent of the June transactions that @kevin posted. The automated trading at IB should ensure that quick (unintended) round tripping won't happen while still firing the entry trade within seconds of the trigger having been reached.

@kevin_in_GA Although using a smaller SD would, in theory, provide for more entry opportunities, I respect the 2.5 SD that you indicated back in June. I just seem to have hit a brick wall in coding PTL to match your PTF findings.

In the end, I have no problem committing sufficient funds to make the $$ amount sufficient as long as the win rate and the average payout are worth the effort. While commissions are always a consideration and they are part of my decision making process, I'm OK with covering the extra event in theory (assuming a resolution is achieved related to the PTL settings).

Not trying to belabor this discussion, but it seems as though there is a solution to be had with this particular pair.

StefanM
25 posts
msg #130563
Ignore StefanM
8/25/2016 2:35:01 AM

@Kevin_in_GA
Kevin, my contribution is not related to this thread, but I read somewhere else that you use StrataSearch. I found a bug there yesterday with the price of one stock after downloading it through Yahoo finance. If you look at NEE then you see that StrataSearch/Yahoo uses wrong prices for this stock.
I wrote the support and they confirmed the issue. If you want then I can send you their answer, possibly I found you at the former BMT website and there I can send you this in private.
However, it does not fix the problem at the moment ..

ferndave
54 posts
msg #130570
Ignore ferndave
8/25/2016 12:13:42 PM

@c1916 that's the theory. In practice, I've watched PTL not execute trades after reaching the set threshold. It executes close, but not exactly. It may be a bug in their software. All I know is that the June results from PTF are completely unrealistic in PTL. I've been running over 30 pairs of mixed linearity for a month with SDs between 1.5-2.0 and not one has had entries and exits like those June numbers. I haven't had more than 4 round trips for a single pair.

c1916
70 posts
msg #130585
Ignore c1916
8/25/2016 7:54:07 PM

Thanks for that Ferndave. It sounds like I may not be as crazy as I initially thought then, as your experience mirrors mine (within reason as I've only seen two round trips).

I wish I knew the specifics behind the algorithm @kevin ran in PTF in June, but the "standard" pairs protocol we were discussing back then (associated with PTL) seems to be a bust.

Kevin_in_GA
4,548 posts
msg #130586
Ignore Kevin_in_GA
8/25/2016 8:06:53 PM

Not trying to mislead anyone - I posted the exact settings for these already, and those are the intraday results provided by PTL, including the timestamps. What I have learned is that these trades could not have been made in reality since they were based on the close and without any slippage in price. Once those are factored in, the trade frequency drops to almost nothing and the trades are hyper sensitive to small changes n price that lead to poor performance.

That is why I have cautioned folks about these pair trades. Make no mistake - I traded VIXY?VXZ in my IB account using PTL and while the trades were triggered as promised off the bid and ask prices, my pair selection was based on data from PTF and never played out as expected.

c1916
70 posts
msg #130587
Ignore c1916
8/25/2016 8:20:15 PM

@kevin. I know that you stated the parameters previously. I wasn't suggesting anything misleading. I assumed that I had not transcribed the settings properly because I was never able to recreate the PTL results (even in test). I assumed operator error on my part, nothing nefarious on your end.

This has been an entertaining exercise. Not ready to give up, but I'll be moving away from the volatility pair for the reasons explained in detail here.

StockFetcher Forums · Filter Exchange · BIGG Brexit Meeting next week...how is everyone playing it?<< 1 ... 3 4 5 6 7 >>Post Follow-up

*** Disclaimer *** StockFetcher.com does not endorse or suggest any of the securities which are returned in any of the searches or filters. They are provided purely for informational and research purposes. StockFetcher.com does not recommend particular securities. StockFetcher.com, Vestyl Software, L.L.C. and involved content providers shall not be liable for any errors or delays in the content, or for any actions taken based on the content.


Copyright 2016 - Vestyl Software L.L.C.Terms of Service | License | Questions or comments? Contact Us
EOD Data sources: DDFPlus & CSI Data Quotes delayed during active market hours. Delay times are at least 15 mins for NASDAQ, 20 mins for NYSE and Amex. Delayed intraday data provided by DDFPlus