volatilityhq.com Official Thread

[Javier 99]

Thanks that what I was doing. It works 100% of the times in an Apple tablet and something like 20% of the times in two different Windows computers, must be click tatency, WiFi speed or ... who knows. Anyway that was my nice to have feature. Thanks for your excellent support.

[Djtux 548]

On 2/22/2019 at 1:01 PM, againstallodds said:

@DjtuxCould you look at BRK.B ? I try to load the straddle graphs but nothing comes up.

BRK.B should work now for straddle RV.

[Djtux 548]

Added support for long stock in the return matrix (the heatmap).

[Djtux 548]

On 2/12/2019 at 9:42 PM, luxmon said:

@Djtux I realize you're probably inundated with requests with this tool, but would it be possible in the scanner to report the total open interest in the weekly expiration series?  

This feature is deployed in production along with the open interest in the first monthly expiry after the earning date.

If there are no weeklies or if the monthly expiry is the same as the weekly expiry, then the open interest is None.

[Djtux 548]

I've added 3 new columns to track the RV rise of the calendars (double, put, call).

RV rise is calculated as the change in % in RV for calendar between the previous day and the current RV. So calendar_price(latest available date) / calendar_price(latest available date - 1) - 1.0.

On 9/2/2018 at 1:49 PM, Arthur said:

I would just look for calendars with the highest RV rise (%) and then apply further restrictions.

@craigsmith might be interested as well.

[luxmon 233]

@Djtux Thank you for the recent enhancements to your tool, they are already proving very helpful to me.

On another topic, could you discuss a bit on how the "Last implied volatility" data is determined?  It seems to not always indicate the IV of the earnings expiration if the earnings is too far out.  For example, see the YELP example below where "L. IV" is reported as ~34% but the May10 term is 56%.  Is there a window of time it looks in to determine the level?

Thanks much,

Tim

[Djtux 548]

40 minutes ago, luxmon said:

@Djtux Thank you for the recent enhancements to your tool, they are already proving very helpful to me.

On another topic, could you discuss a bit on how the "Last implied volatility" data is determined?  It seems to not always indicate the IV of the earnings expiration if the earnings is too far out.  For example, see the YELP example below where "L. IV" is reported as ~34% but the May10 term is 56%.  Is there a window of time it looks in to determine the level?

Thanks much,

Tim

Thank you for the encouragement. The IV is 30 calendar day implied volatility calculated a little bit like the VIX.

[luxmon 233]

31 minutes ago, Djtux said:

Thank you for the encouragement. The IV is 30 calendar day implied volatility calculated a little bit like the VIX.

Thank you for clarifying!  I'm as reluctant as anyone to suggest adding more columns to the scanner (especially after just updating my Excel VB scripts to point to shifted column references), but.. I must say having the earnings expiration IV (projected or confirmed) reported would be icing on the cake to help scan for interesting pre-pre earnings calendars if it could be squeezed into a future update 😉

[Djtux 548]

On 3/29/2019 at 9:56 PM, luxmon said:

Thank you for clarifying!  I'm as reluctant as anyone to suggest adding more columns to the scanner (especially after just updating my Excel VB scripts to point to shifted column references), but.. I must say having the earnings expiration IV (projected or confirmed) reported would be icing on the cake to help scan for interesting pre-pre earnings calendars if it could be squeezed into a future update 😉

I've added a new column E.IV in the scanner which represents the IV of the atm options with the first expiry date after the earnings date.

[luxmon 233]

18 minutes ago, Djtux said:

I've added a new column E.IV in the scanner which represents the IV of the atm options with the first expiry date after the earnings date.

Most excellent!  Having this raw data opens up many more avenues for filtering trade candidates.

Many thanks.

[Javier 99]

As You know We are currently using the Call Return Matrix (Deltas 80,60,40) to determine the potential of Bull Ratio spreads (TrustyJules style). The hot map spots are too biased by outliers. Is it feasible to ignore outliers (as we have in many other charts)? (i.e. Remove cycles with hundreds or even thousands per cent performance).

[Djtux 548]

3 minutes ago, Javier said:

As You know We are currently using the Call Return Matrix (Deltas 80,60,40) to determine the potential of Bull Ratio spreads (TrustyJules style). The hot map spots are too biased by outliers. Is it feasible to ignore outliers (as we have in many other charts)? (i.e. Remove cycles with hundreds or even thousands per cent performance).

Can you give me a ticker you have in mind ?

Normally if you use the median, it should be less sensitive to outliers.

[Yowster 9,174]

46 minutes ago, Javier said:

As You know We are currently using the Call Return Matrix (Deltas 80,60,40) to determine the potential of Bull Ratio spreads (TrustyJules style). The hot map spots are too biased by outliers. Is it feasible to ignore outliers (as we have in many other charts)? (i.e. Remove cycles with hundreds or even thousands per cent performance).

Those charts use the median, not the average... so the outlier effect is less than if the average was used.     Personally, I like to look at the straddle chart with the "show stock price change" enabled in the Advanced Options - the chart that this produces gives you a nice visual graph of each cycle so you can quickly see the stock price rising cycles compared to the falling cycles (for the TrustyJules ratio trades you ideally want to see the majority of cycles showing the stock price rise).

[TrustyJules 3,212]

What I am working on at the moment is seeing how to run all the different deltas outcomes against each other. That should give a better idea of good entry and exit times. Empirically its clear also that nearer to earnings works a little different than closer to earnings.The issue of the outliers is fairly easy to avoid you just filter out when more than 1/3 of the returns deviate in outcome from the others - ideally you want the most consistent return to work with.

 

[Javier 99]

2 hours ago, Djtux said:

Can you give me a ticker you have in mind ?

Normally if you use the median, it should be less sensitive to outliers.

 

1 hour ago, Yowster said:

Those charts use the median, not the average... so the outlier effect is less than if the average was used.     Personally, I like to look at the straddle chart with the "show stock price change" enabled in the Advanced Options - the chart that this produces gives you a nice visual graph of each cycle so you can quickly see the stock price rising cycles compared to the falling cycles (for the TrustyJules ratio trades you ideally want to see the majority of cycles showing the stock price rise).

 

1 hour ago, TrustyJules said:

What I am working on at the moment is seeing how to run all the different deltas outcomes against each other. That should give a better idea of good entry and exit times. Empirically its clear also that nearer to earnings works a little different than closer to earnings.The issue of the outliers is fairly easy to avoid you just filter out when more than 1/3 of the returns deviate in outcome from the others - ideally you want the most consistent return to work with.

 

Thanks to you all three, agree with median less impacted than average by outliers of course, but still don’t see why not discard outliers as in other charts if not technically too complex. I’ll go in depth with your posts and maybe come back. Thanks again.

[TrustyJules 3,212]

6 hours ago, Javier said:

Thanks to you all three, agree with median less impacted than average by outliers of course, but still don’t see why not discard outliers as in other charts if not technically too complex. I’ll go in depth with your posts and maybe come back. Thanks again

Why not? Because they occur - furthermore the past is no guarantee for the future. My approach to the heatmaps is just to narrow the probabilities slightly in our favour and combine this with the known rise of IV to open a position where we have market exposure and more than a regular chance of getting directional bias in our favour. Simultaneously the downside is protected by the rise in IV.

When looking at candidates the most important thing is whether the return is regularly positive - is there is a 1 in 2 chance of 400% + and otherwise 10% - that is not good. If there is a 1 in 2 chance of 50% or -25% that is bad. If the chance is 3 to 1 in favour of the positive then again it gets interesting.

[luxmon 233]

I noticed when backtesting calendars where the short expiration expires before earnings with 1-week spacing and the earnings date is relatively near (I haven't quantified how near), the tool seems to not be able to compute the calendar for the current cycle.  I've attached two examples of the error messages where I'm backtesting for the Apr12/18 calendars and the option series that cannot be found.  Are there some settings that need to be tweaked in this case?

Updated:  Maybe it's due to the Good Friday holiday on 4/19, and the options that week expire on Thursday(?)

Thanks

AXP:

UNP:

Edited April 6, 2019 by luxmon
Added note about Good Friday market holiday

[Djtux 548]

2 hours ago, luxmon said:

I noticed when backtesting calendars where the short expiration expires before earnings with 1-week spacing and the earnings date is relatively near (I haven't quantified how near), the tool seems to not be able to compute the calendar for the current cycle.  I've attached two examples of the error messages where I'm backtesting for the Apr12/18 calendars and the option series that cannot be found.  Are there some settings that need to be tweaked in this case?

Updated:  Maybe it's due to the Good Friday holiday on 4/19, and the options that week expire on Thursday(?)

Thanks

• Thank you for the 2 examples, that helped the investigation of the issue. It's a bug in the code due to Good Friday, you are right.

I just fixed it, but there are some delays for the import of the data from yesterday.

• I've added a new column in the "Straddle Table' page (bad name) : historical earnings volatility which calculates the historical volatility of the 8 previous stock movement around the earning. It's still a work in progress.
• I've been doing some changes to display the strikes in the RV chart when you hover on a point, but it will only appear when i reset the RV charts cache. I will probably do that during Good Friday long weekend to have the time to rebuild the whole cache.

[luxmon 233]

@Djtux I've been checking periodically but it appears today's data is not yet available in the scanner.  Do you think it will be up tonight?

Thanks

[Djtux 548]

6 minutes ago, luxmon said:

@Djtux I've been checking periodically but it appears today's data is not yet available in the scanner.  Do you think it will be up tonight?

Thanks

The import is still running. The data provider was late today and lately the import process is getting just slower because of the all the calculations that needs to be done for the scanner and recent 'new' features. I will work in the next few weeks to see what can be done to improve the import time. Sorry about that. 

[luxmon 233]

1 minute ago, Djtux said:

The import is still running. The data provider was late today and lately the import process is getting just slower because of the all the calculations that needs to be done for the scanner and recent 'new' features. I will work in the next few weeks to see what can be done to improve the import time. Sorry about that.

No worries, I've become spoiled as it's almost always there by the time I run the scan in the evening!  The new features are great so I'll take the side effects once in a while.

[luxmon 233]

AMAT is being reported as having a confirmed earnings date by two of the sources, but they have not confirmed on their website (and usually don't until the first week in May).  Curiously enough, yesterday only source #4 was showing confirmed.  I suspect these sources have their wires crossed?

 

[Djtux 548]

39 minutes ago, luxmon said:

No worries, I've become spoiled as it's almost always there by the time I run the scan in the evening!  The new features are great so I'll take the side effects once in a while.

You were not the only person impacted, that's why. My previous message was just to state that i take performance, downtime and fast data import seriously. I'm not sure how i'm going to improve it but i will try.

[Djtux 548]

25 minutes ago, luxmon said:

AMAT is being reported as having a confirmed earnings date by two of the sources, but they have not confirmed on their website (and usually don't until the first week in May).  Curiously enough, yesterday only source #4 was showing confirmed.  I suspect these sources have their wires crossed?

 

No i'm pretty sure they don't use the same source because often they don't agree.

For AMAT, i don't know the stock but there is this date : http://investors.appliedmaterials.com/phoenix.zhtml?c=112059&p=irol-calendar

I don't know if they change their date.

[zxcv64 771]

@djtux, I've searched for documentation on the following feature but couldn't find it. So, maybe you can pls clarify :

 

1) Under the 'Straddle' top level drop-down, there is an option called 'Table' which shows a histogram and a table of various things including the below

 

Can I ask if the 'Straddle return' is return for a straddle that is bought at T-0 and closed on the first full day of the earnings? Eg. for HAS, earnings are BMO Tue 23-Apr, so would the return be calculated as [(straddle price on Tue 23-Apr at market close)  - (straddle price on Mon 22-Apr at market close)/(straddle price on Mon 22-Apr at market close) ?

2) For the "Call Calendar Return", would it be similar to above, with the short calls expiring the first Fri after earnings and the longs expiring on the next monthly date? For HAS, the shorts would be Fri 26-Apr and the longs 17-May?

3) I'm curious how the max loss on a Call Calendar Return can be greater than 100% (see above)?  Also, I think the Dbl Calendar Return is incorrect. (Is the Dbl Calendar Return a combination of the Call and Put Calendar Returns?)

 

[Djtux 548]

Quote

Eg. for HAS, earnings are BMO Tue 23-Apr, so would the return be calculated as [(straddle price on Tue 23-Apr at market close)  - (straddle price on Mon 22-Apr at market close)/(straddle price on Mon 22-Apr at market close) ?

Yes that is correct.

Quote

For the "Call Calendar Return", would it be similar to above, with the short calls expiring the first Fri after earnings and the longs expiring on the next monthly date? For HAS, the shorts would be Fri 26-Apr and the longs 17-May?

Correct, short the first weekly (if there is a weekly) and leg the first monthly. Both are after the earnings date.

Quote

I'm curious how the max loss on a Call Calendar Return can be greater than 100% (see above)?  Also, I think the Dbl Calendar Return is incorrect. (Is the Dbl Calendar Return a combination of the Call and Put Calendar Returns?)

I think in this case, the problem comes from the bid-ask spread which is large, and i'm basing all my calculations on the mid price, which is ok for liquid stocks, but not realistic for illiquid stock. I only have the bid/ask prices of individual options, not spreads, so it's difficult to know how much that calendar could get filled.

I'm double checking with ThinkOrSwim Thinkback.

The call calendar on Feb 06 2018 is at 0.10 credit because of the bid/ask spread, and i think that's why the -250% comes from. Return is (CalendarPriceT+1/CalendarPriceT-0) -1

The next day, it's 0.15 debit.

 

The double calendar is just a call calendar + put calendar.

Return is (0.275/0.05-1)=450%. Obviously irrealistic because of the bid/ask spread and commissions.

 

I think i should at least implement some sort of a tooltip so that when you put your mouse on the return, you get that information to decide why it's like that.

 

[luxmon 233]

@Djtux FYI, for some reason the weeklies data for CSCO is not showing up in the Scanner output.

[Djtux 548]

3 minutes ago, luxmon said:

@Djtux FYI, for some reason the weeklies data for CSCO is not showing up in the Scanner output.

Yes this is a change i made few weeks ago when i added some other feature and i saw that issue where the first expiry after the earning date was the monthly. I decided to set the weeklies numbers like the volume, OI, spread at None because that expiry is actually a monthly, not weekly.

I'm not sure what's the best way to deal with it :

• keep it as is : None if the first expiry after earning is actually a monthly 
• Change the message from None to something else to say that the weekly doesn't exist
• Copy the value from the monthly over, but it's a little bit misleading

Any opinion ?

[luxmon 233]

11 hours ago, Djtux said:

I'm not sure what's the best way to deal with it :

• keep it as is : None if the first expiry after earning is actually a monthly 
• Change the message from None to something else to say that the weekly doesn't exist
• Copy the value from the monthly over, but it's a little bit misleading

Any opinion ?

Thanks for the explanation, and I see the conundrum. After thinking about it, I would say it's not a big deal for the Scanner to say None since you have the "W" field indicating if weeklies are available (allowing one to still accurately filter for 1 or 2-week eligible calendar candidates). 

Thanks,

Tim 

[Arthur 32]

@Djtux Would it be possible to add a new column to the Scanner: Average Max Move from the last 4 cycles?

For the background: I sometimes hold straddles through earnings hoping for a large stock move. Being able to compare the current straddle RV to the average max move from the last 4 cycles would make it much easier for me to find suitable through-earnings-candidates. Currently I have to do the calculations manually for each stock using the Straddle Table.

 

[Yowster 9,174]

25 minutes ago, Arthur said:

@Djtux Would it be possible to add a new column to the Scanner: Average Max Move from the last 4 cycles?

For the background: I sometimes hold straddles through earnings hoping for a large stock move. Being able to compare the current straddle RV to the average max move from the last 4 cycles would make it much easier for me to find suitable through-earnings-candidates. Currently I have to do the calculations manually for each stock using the Straddle Table.

 

It might not be exactly what you are looking for, but the straddle RV chart visually shows the average 1-day earnings move (green dotted line) and the max 1-day earnings move (dotted red line).

[Djtux 548]

7 hours ago, Arthur said:

@Djtux Would it be possible to add a new column to the Scanner: Average Max Move from the last 4 cycles?

For the background: I sometimes hold straddles through earnings hoping for a large stock move. Being able to compare the current straddle RV to the average max move from the last 4 cycles would make it much easier for me to find suitable through-earnings-candidates. Currently I have to do the calculations manually for each stock using the Straddle Table.

 

I will put into my todo list. There will be 2 new columns : average max move and average price effect for the last X years (with X controlled by the user and can be 1, 2 or 3 years).

[Djtux 548]

On 4/15/2019 at 10:25 PM, Djtux said:

The import is still running. The data provider was late today and lately the import process is getting just slower because of the all the calculations that needs to be done for the scanner and recent 'new' features. I will work in the next few weeks to see what can be done to improve the import time. Sorry about that. 

There is a delay in the data import today even though i did some improvements in the past few weeks. I still have few ideas to implement to prevent that.

[speedythewarthog 32]

Hi @Djtux, just subscribed to your web service and enjoying looking around at all the features. I noticed under beta features you have a call ratio calculator, which generates a return matrix for a given pair of call deltas, and produces a ratio.

Can you please explain how the %return is calculated, and what the ratio means? How much historical data feeds into these?

Thanks!

-speedy

[Djtux 548]

20 minutes ago, speedythewarthog said:

Hi @Djtux, just subscribed to your web service and enjoying looking around at all the features. I noticed under beta features you have a call ratio calculator, which generates a return matrix for a given pair of call deltas, and produces a ratio.

Can you please explain how the %return is calculated, and what the ratio means? How much historical data feeds into these?

Thanks!

-speedy

Welcome aboard !

 

There are some links here : https://steadyoptions.com/forums/forum/topic/5432-john-de-ratio/?do=findComment&comment=125118. The posts are not in the right thread, sorry about that.

I'm still working on the call ratio calculator, but it's available in beta to get feedbacks.

Quote

 

Yes it's 8 cycles. The increase column is the same as the cells in C8:F11. It was for me to double check the numbers against @TrustyJules spreadsheet.

https://steadyoptions.com/forums/forum/topic/5168-the-rational-trader/?do=findComment&comment=118665

 

Quote

 

The percentage return is the matrix C13:F16 in the spreadsheet https://steadyoptions.com/forums/forum/topic/5168-the-rational-trader/?do=findComment&comment=118665

I'm basically automating the spreadsheet by retrieving for each delta call, the strike, the price of that option yesterday, and the % median increase from the 'return matrix' heatmap.

I'm still working on it, and looking for feedbacks on how to improve it.

 

So to answer your questions : the % return and the ratio are taken from @TrustyJules spreadsheet that is available in the Rational Trader thread, and it's using 8 earnings cycles for now.

Edited May 20, 2019 by Djtux
typos

[Javier 99]

@Djtux Cosmetic question: How difficult is to add some reference dates, together with T-x points in the graphs?. Let´s say, for instance, every Friday date (5/31/2019 ... 6/7/2019).

I use to check my computer calendar back and forth to figure out what T-5 date stands for. 

[Djtux 548]

3 minutes ago, Javier said:

@Djtux Cosmetic question: How difficult is to add some reference dates, together with T-x points in the graphs?. Let´s say, for instance, every Friday date (5/31/2019 ... 6/7/2019).

I use to check my computer calendar back and forth to figure out what T-5 date stands for. 

It's already in the chart i believe.

You can look for the marker at the bottom to see the fridays (the bold is the monthly) : see the 2 red arrows on the left.

On the average hover text, you will also see the date '2019-06-07 Friday' corresponding to the T-8 for example.

[Javier 99]

9 hours ago, Djtux said:

It's already in the chart i believe.

You can look for the marker at the bottom to see the fridays (the bold is the monthly) : see the 2 red arrows on the left.

On the average hover text, you will also see the date '2019-06-07 Friday' corresponding to the T-8 for example.

Fine, didn’t notice. Something similar for the return matrix chart?

[Djtux 548]

On 5/30/2019 at 2:35 AM, Javier said:

Fine, didn’t notice. Something similar for the return matrix chart?

You are right, adding the actual date in the return matrix would help.

[Djtux 548]

On 6/22/2019 at 9:08 AM, Djtux said:

It’s going to be a desktop application that you will download from my website and install on your machine. The software on your machine could then connect to your broker software to get the realtime data.  For now i will start with IB.

When i have a beta i will put a message on the official thread found in my signature.

There is a new page under the 'Beta' menu to download the app. It's not even a beta but rather an alpha and very basic. It's only for the most adventurous.

[Djtux 548]

On 5/30/2019 at 2:35 AM, Javier said:

Fine, didn’t notice. Something similar for the return matrix chart?

The entry actual date and exit actual date should be visible in the return matrix (the heatmap) right now.

[Javier 99]

22 hours ago, Djtux said:

There is a new page under the 'Beta' menu to download the app. It's not even a beta but rather an alpha and very basic. It's only for the most adventurous.

Some initial issues/questions, maybe a bit naif ones:

1) As I use IB as online data feeder for One, I needed to change One Port to 7496 as well to keep working with both apps.

2) I cannot  see any log tab, so I didn´t know if your app is working properly or stopped. Seems it is too big to fit in my computer screen. Do You know how to reduce the size (tried with traditional "Ctrl -" with no result).

3) What straddle/calendar are You meaning?. The one just after earnings date?

 

[Djtux 548]

11 minutes ago, Javier said:

Some initial issues/questions, maybe a bit naif ones:

1) As I use IB as online data feeder for One, I needed to change One Port to 7496 as well to keep working with both apps.

2) I cannot  see any log tab, so I didn´t know if your app is working properly or stopped. Seems it is too big to fit in my computer screen. Do You know how to reduce the size (tried with traditional "Ctrl -" with no result).

3) What straddle/calendar are You meaning?. The one just after earnings date?

 

1. Noted. At some point, the hostname, ports and clientid should be configurable too.
2. Are you able to resize the window by using the mouse in one corner of the window ? It will reduce the default size in a future release.
3. The straddle has the first expiry after the earning date. For the calendar, the short leg is the first expiry after the earning date, and the long leg is the first monthly expiry after the short leg (so both legs are after earnings).

[Javier 99]

1 hour ago, Djtux said:

1. Noted. At some point, the hostname, ports and clientid should be configurable too.
2. Are you able to resize the window by using the mouse in one corner of the window ? It will reduce the default size in a future release.
3. The straddle has the first expiry after the earning date. For the calendar, the short leg is the first expiry after the earning date, and the long leg is the first monthly expiry after the short leg (so both legs are after earnings).

Ok, great, first of all thank You for this toy, I will make an intensive use of it. I solved all my small issues, my comments to your comments:

1) I set One port to 7496 as well and both, One and your app, seem to be working smoothly. Solved

2) No way to resize using the mouse, all corners were out of the screen, but "ALT + space bar" let me resize to Full Screen. Solved.

3) I figured out You were using Straddle and Calendar as in the rest of your Web page (next expiration after earnings + following monthly exp for the long leg of the calendar). Solved

[Djtux 548]

If you close the desktop app, and relaunch it, it will update automatically.

You will be able to test the watchlist.

Again, it's still buggy and slow.

[Javier 99]

Using your tool Today. What does it mean these X-axis funny numbers?

[Yowster 9,174]

1 hour ago, Javier said:

Using your tool Today. What does it mean these X-axis funny numbers?

 

X-axis looks to be days to earnings, so you see the fractional amount during the day and the whole number 9 at the close (I think @Djtux said it was off by 1 day, hence the 9 instead of 10).

[Djtux 548]

1 hour ago, Javier said:

Using your tool Today. What does it mean these X-axis funny numbers?

 

18 minutes ago, Yowster said:

X-axis looks to be days to earnings, so you see the fractional amount during the day and the whole number 9 at the close (I think @Djtux said it was off by 1 day, hence the 9 instead of 10).

Yes it's supposed to be the number of trading days before the earning date. It's fractional so that -9.5 is the middle of the day for example. -9.9 is near the open for that day, and -9.01 is near the close.

Right now, there is a bug in the counting of the number of days for the x-axis to exclude the weekends and the holiday (the 4th of july), but your chart ASML shows the straddle RV for ASML for June 28th (today), and -9.987 is the open for today, and -9.01 is the close for today. The number 9 is wrong and off by a few days, but you get the idea.

[Djtux 548]

@Javier

Here is the fixed ASML x-axis.

My goal for the x-axis in the desktop app is to replicate the same x-axis found in the RV charts on the website, so that it's easier to compare.

See the green arrow for the trading session of June 27th, and the red arrow for the trading session of June 26th.

Also note that from the IB TWS api, you don't get the last close of 16:00, so if you look at a 1 min chart, the last available point from the api is 15:59. A little bit silly but it seems a limitation of the api.

 

[mustafaoe 10]

@DjtuxIs there a documentation of the web services you provide? I am asking myself if it is possible to read the scanner results daily into a program?