Using the IQfeed history Viewer, the following options show:

Wednesday expiration, April 2022, Put, strike 4,200.00:

@E1CJ22P420000 (week=1, data through 4/06/2022)
@E2CJ22P420000 (week=2, data through 4/13/2022)
@E3CJ22P420000 (week=3, data through 4/20/2022)
@E4CJ22P420000 (week=4, data through 4/27/2022)

April 2022, end of week options:

@EW1J22P420000 (week=1, data through 3/31/2022)
@EW2J22P420000 (week=2, data through 4/08/2022)
@EW3J22P420000 (week=3, data through 4/14/2022)
@EW4J22P420000 (week=4, data through 4/22/2022)
@EW5J22P420000 (week=5, no data)
@EWJ22P420000 (EOM, data through 4/29/2022)
(EOM coincides with end of last week for April 2022)

What is the algorithm that defines the starting day and ending day of the “week of month” numbers, for the cases of:

Case 1: @E1*, @E2*, @E3*, @E4*, @E5*
and
Case 2: @EW1*, @EW2*, @EW3*, @EW4*, @EW5*?

It seems that the week number continues into the following calendar week in Case 1 (@E1*, Etc.)?

Thanks. Any pointers to documentation on the “week of month” algorithms are appreciated.
Edited by step999 on May 6, 2022 at 06:22 AM

I'll confirm this new nomenclature, but I'm pretty sure you've got it right.

---

Sincerely,
Gary Stephen
DTN IQFeed Implementation Support Specialist

Hi DTN Gary Stephen,

My concern is with the new @ES option expiries (an expiry every day), I can’t go back in time to check out a hypothesis of what algorithm DTN is using.

In any event, knowing your exact “week of month” algorithm would be much more reliable than trying to guess the DTN rules (from years of back month data), especially with the effects of holidays.

For example, if we assume we have in the “first” week @E1A, @E1B, @E1C, @E1D (for Mon, Tue, Wed, Thu expirations) . . . . Must that “first” week always be contained inside one calendar week?

What if the first calendar week only has Wednesday & Thursday expiries? Then is the next Monday expiry named @E2A?

Or is some other algorithm used? In addition, how do holidays affect the naming in your DTN naming algorithm for @ES futures options?

DTN must have some algorithm that codifies this naming.

Thanks for your reply.
Edited by step999 on May 6, 2022 at 04:17 PM

Generally, the new weekly futures options for futures will follow this pattern:

A = Monday expiration
B = Tuesday expiration
C = Wednesday expiration
D = Thursday expiration

Friday may or may not be E. Before expirations on different days of the week were introduced, the default was that equity options expired on Friday, so it didn't need anything to indicate the weekday. Some of these symbols have been grandfathered in.

Furthermore, the ABCD letter indicating the weekday of the expiring option might not appear in the same place for different future options.

It is coincidental that EW options expire at the end of the week. They were the first weekly options introduced on the E-mini S&P 500. The symbol logic at the time was most likely E = E-mini, W = weekly.

I infer from the names of past future options that the week number is limited to the month it actually belongs to:

LM,@Q5CH22P995000,"E-MINI NASDAQ 100 WEDNESDAY WEEK 5 MARCH 2022 PUT 995000",,,,,,,,,,,2,12,,,2.24,43,,"@NQM22",,,,,,,,,,,,,,,,,,,,,20220330,,,,,,

LM,@QN1J22C1000000,"E-MINI NASDAQ 100 WEEK 1 APRIL 2022 CALL 1000000",,,,,,,,,,,2,12,,,2.33,43,,"@NQM22",,,,,,,,,,,,,,,,,,,,,20220401,,,,,,

Future Options expiring on Wednesday March 30 were "WEDNESDAY WEEK 5 MARCH" but ending on Thursday, April 1 were "WEEK 1 APRIL".

This may be applied inconsistently, though.

---

Sincerely,
Gary Stephen
DTN IQFeed Implementation Support Specialist

Hi DTN Gary Stephen,

We agree on the “day of week” nomenclature for A=Monday, B=Tuesday,
C=Wednesday, D=Thursday expirations.

I have not yet seen a “day of week” E=Friday designation, but I understand that
would be logical. As you say, it seems customary instead to use the
grandfathered-in W* (W1,W2,etc.) designations for Fridays.

So I think we agree on the above “day of week” nomenclatures.

However, my basic question is about the “week of month” numbers
(I will use as shorthand wN).
These wN are more difficult to understand.

One way to look at this is to understand the differences between:
1) Calendar “week of month” numbers
2) Futures options A,B,C,D “week of month” numbers, or wN
3) Futures options W* (Friday) “week of month” numbers

Assuming an expiration can only occur on a weekday (Monday through Friday) . . . .

The rules that determine a calendar “week of month” number (wN)
(not necessarily used in trading):

1) The week number is calculated only from days within the calendar month
2) The first calendar “week of month” number (wN) 
     begins on the earliest day of that month, 
      that day also being a weekday
3) The wN numbers for the month are incremented at each new Monday

However, for the rules for futures options A,B,C,D, the
“week of month” numbers (wN) become more complicated:

1) The week number is calculated only from days within the calendar month
2) The first calendar “week of month” number (wN)
     begins on the earliest day of that month, 
     that day also being a weekday,
	  AND
     there is at least one valid A,B,C,D expiration day in
     that calendar week
     (else the week is not counted for A,B,C,D  wN’s)
3) The wN numbers for the month are incremented at each new Monday
    (for both A,B,C,D numbers, and W* numbers)

Example:  there is an A,B,C,D expiration in the first calendar week . . . . 

             S   M   T   W   T   F   S   S   M  (Sun, Mon, etc.)
                1A  1B  1C  1D  W1
 DayOfMonth= 1   2   3   4   5   6   7   8   9
                 ^--start of nW=1
                                             ^--start of nW=2

Note:  in the above case, the “week of month” number for Friday is
the same as the “week of month” number for A,B,C,D

Example:  no A,B,C,D expiration in the first calendar week . . . .

             F   S   S   M   T   W   T   F   S   S   M  (Fri, Sat, etc.)
            W1          1A  1B  1C  1D  W2
 DayOfMonth= 1   2   3   4   5   6   7   8   9  10  11 
             ^--NOT nW=1
                         ^--start of nW=1
                                                     ^--start of nW=2
Note:  in the above case, the “week of month” number for the second
Friday (W2) is different than “week of month” number
for A,B,C,D (1A,1B,1C,1D) in that same calendar week.  
This is why some months only have four (4) A,B,C,D  “week of month” 
numbers, but have five (5) calendar weeks, and possibly five (5) W*
numbers (Fridays).

For holidays, I believe the holidays would be treated as if they were
NOT a weekday (would not count as a weekday day for counting wN).
Please correct if I am mistaken.

Again, I wish to find the exact rules or algorithm for the:
- - Futures options A,B,C,D “week of month” numbers
- - Futures options W* (Friday) “week of month” numbers

I believe your DTN programmers would know all the rules,
since they coded them

Thanks.

DTN doesn't make the symbols - the exchange does. DTN endeavors to use the same symbol as the exchange whenever possible. When conflicts with another exchange's symbols occur, there aren't any strict rules, or algorithm for how to change it, except that we try to match the exchange's symbol as much as possible. Typically, we just precede it with @ or Q.

I wasn't able to find any official rules for how CME numbers the weeks for weekly options. I inferred that Week 1 starts on the 1st day of the month, with no regard to what day of the week it is. But I only looked at one particular month to determine that. This page https://www.cmegroup.com/trading/options/weekly-options-on-futures.html is the CME's list of Product Codes for future options. Interestingly, there's no code for "5th week end of the month" for E-mini S&P 500 - only through Thursday.



A 5th Friday of the month can exist, so I'm not sure what this implies. Perhaps the CME handles week counting differently for "day-of-the-week" future options than it does for "end of week" future options. (Bearing in mind what I said earlier about Friday not getting its own code because "end of week" covered it.) I'll see if I can learn anything from the CME; I didn't find their naming rules online, like it's easy to find the OSI standard for equity options online.

---

Sincerely,
Gary Stephen
DTN IQFeed Implementation Support Specialist

Hi DTN Gary Stephen,

Thanks for explaining that your DTN code mirrors the option names as
determined by the CME.

I apologize for assuming that the DTN code had logic that duplicated the CME
naming rules.

Like you, I was unable to find the CME futures options naming rules for
the “week of month” number on any web page or document, CME or otherwise.

As you mention:
Perhaps the CME handles week counting differently for "day-of-the-week" future
options than it does for "end of week" future options.

I believe this is the case. Please see my example from my first post in this topic above,
dated May 6 and repeated here…. (The option names are valid IQfeed names.)

Data from April 2022:

Wednesday expiration, April 2022, Put, strike 4,200.00:

@E1CJ22P420000   (week=1, data through 4/06/2022)
@E2CJ22P420000   (week=2, data through 4/13/2022)
@E3CJ22P420000   (week=3, data through 4/20/2022)
@E4CJ22P420000   (week=4, data through 4/27/2022)

April 2022, end of week options:

@EW1J22P420000   (week=1, data through 3/31/2022)
@EW2J22P420000   (week=2, data through 4/08/2022)
@EW3J22P420000   (week=3, data through 4/14/2022)
@EW4J22P420000   (week=4, data through 4/22/2022)
@EW5J22P420000   (week=5, no data)
@EWJ22P420000    (EOM, data through 4/29/2022)
 (EOM coincides with end of last week for April 2022)

Notice that the 1C,2C,3C,4C (Wednesday expiry) options for the same
“week of month” number expire in the calendar week following the
W1,W2,W3,W4 (Friday) options.

So it must be the case that the "week of month" for A,B,C,D (Mon,Tue,Wed,Thu) options
is not always the same as the "week of month" for the W1,W2,W3,W4 (Friday)
options - and therefore they use different naming rules.

I gave an illustration of this in my last post:

Example:  no A,B,C,D expiration in the first calendar week . . . .

             F   S   S   M   T   W   T   F   S   S   M  (Fri, Sat, etc.)
            W1          1A  1B  1C  1D  W2
 DayOfMonth= 1   2   3   4   5   6   7   8   9  10  11 
             ^--NOT nW=1
                         ^--start of nW=1
                                                     ^--start of nW=2
Note:  in the above case, the “week of month” number for the second
Friday (W2) is different than “week of month” number
for A,B,C,D (1A,1B,1C,1D) in that same calendar week.  
This is why some months only have four (4) A,B,C,D  “week of month” 
numbers, but have five (5) calendar weeks, and possibly five (5) W*
numbers (Fridays).

Thanks for the CME Product Codes link. As far as the CME naming, I believe
they limit the “week of month” for Friday to EW1-EW4 because
the 5th Friday always coincides with the trading end-of-month (EOM) day,
which supersedes (replaces) the EW naming for that Friday. (Either that,
or it could be a holiday.)

Some example months showing a 5th Friday that is the trading EOM are:
2021: April, July, October, December (for January, the 1st Friday is a holiday)
2022: April, July, September, December
(note: The months not listed have 4 Fridays.)

Thanks for your thoughtful replies and information.

I will attempt to reverse engineer and code the rules, and see how they match
to the A,C or A,B,C,D options for the past several months.

Thanks again.

Edited by step999 on May 11, 2022 at 05:03 AM

If I discover anything that will help you I'll post it here. But so far we don't have a clear-cut set of rules for how these new weekday future option symbols are made. There is also room for inconsistency in different underlying commodities, and different ways they handle the legacy symbols.

If you're going to try and reverse-engineer the rules, I suggest using the FDS Market Summary report. This report contains fields for the symbol, full name, and expiration date, like this:

LM,Symbol,Description,PeRatio,AvgVolume,DivYield,DivAmount,DivRate,PayDate,ExDivDate,CurrentEps,EstEps,SIC,Precision,Display,GrowthPercent,FiscalYearEnd,Volatility,ListedMarket,MaturityDate,OptionRoots,CouponRate,InstitutionalPercent,YearEndClose,Beta,LEAPs,WRAPs,Assets,Liabilities,BalanceSheetDate,LongTermDebt,CommonSharesOutstanding,MarketCap,52WeekHigh,52WeekHighDate,52WeekLow,52WeekLowDate,CalHigh,CalHighDate,CalLow,CalLowDate,Expiration,LastSplit,LastSplitDate,PrevSplit,PrevSplitDate,NAICS,ShortInterest
LM,@E1AK22C210000,"E-MINI S&P 500 MONDAY WEEK 1 MAY 2022 CALL 210000",,,,,,,,,,,2,12,,,4.07,43,,,,,,,,,,,,,,,,,,,,,,,20220502,,,,,,

This makes to easy to collect all symbols at once. The command would be FDS,9,34,date in YYYYMMDD format.

---

Sincerely,
Gary Stephen
DTN IQFeed Implementation Support Specialist

Hi DTN Gary Stephen,

Thanks for all of your help on this.

Also thanks for the pointer to the FDS (Fundamental Summary) market
summary command, and the sample syntax.

For reverse-engineering the naming, yes it is quite helpful to see every option
available. (Quite a large number of options, as you are aware.)

The FDS data is revealing some additional patterns/rules.

Anyway, let me work through this and see what naming rules fit all these cases.
They are bit tricky, but the complete rules are becoming more comprehensible.

Thanks very much for the FDS tip, and of course thanks for your other
information and comments as well.

Edited by step999 on May 12, 2022 at 03:44 AM

Thank you, I'm glad to help! And if I may make one last suggestion: don't assume rules are the same for all underlying symbols. We've noticed a little inconsistency in how legacy symbols are handled and so forth.

---

Sincerely,
Gary Stephen
DTN IQFeed Implementation Support Specialist

Hi DTN Gary Stephen,

OK – good to know, so I will check the NQ naming separately for example.

One interesting observation: it seems different brokers can use different rules or
conventions, since it seems TOS (ThinkOrSwim) uses what are basically the calendar
“week of month” for ES futures options, and IB (Interactive Brokers) uses naming
similar to IQFeed/CME for “week of month” numbers.

Also, from first glance using FDS, it seems that the IQfeed/CME convention is to
count the ”week of month” for each week day separately (ignoring all the other
days of the week).

Therefore, the 1A (week 1 Monday futures options) may start at a different calendar
week than the 1C (Wednesday futures options). An example of this is March 2022,
where the first 1C occurs in the first calendar week, but the first 1A occurs in the
following calendar week.

I will check more thoroughly, but if this is true, the algorithm would be simpler
to code.

Thanks very much.

Edited by step999 on May 12, 2022 at 11:04 PM

The CME complete futures options prefix codes are listed here:

https://www.cmegroup.com/education/brochures-and-handbooks/cme-equity-index-options-on-futures-bloomberg-cheat-sheet.html

(not just Bloomberg)
Edited by step999 on May 16, 2022 at 02:35 AM