An observation of TOS after hours when underlying active, but not the options market.

[garyw]

A friend was questioning his SPY option position last night in TOS, which helped to observe this case.
Beware if you reference TOS option information after trading hours IFF the underlying trades after option trading ceases.
SPY, QQQ, TLT are three cases.
User beware... (IMHO TOS is NOT to blame)

Many things will be in error after option trading ceases if the underlying price is disassociated with the options (such as these cases). Extrinsic/Intrinsic value, Theta , ...

 

[status1]

I completely agree with that
The T+0 line can be all over the place even above the expiration line because some of the option prices are either missing or way off
Not sure if there is any way to fix that other than just wait for the market to open but that means you cannot do too much analyzing of the positions It doesn't happen all the time it's mostly when there is a lot of price swings in high volatility

 

[garyw]

I completely agree with that
The T+0 line can be all over the place even above the expiration line because some of the option prices are either missing or way off
Not sure if there is any way to fix that other than just wait for the market to open but that means you cannot do too much analyzing of the positions It doesn't happen all the time it's mostly when there is a lot of price swings in high volatility

Perhaps TOS could be persuaded to include a setting allowing user to specify analysis to reflect valid evaluation times only? -- IE, after hours to use the closing underlying price instead of the "current Mark"! This may also aid in the cases that the BID and ASK vary after market (even though the underlying is not being traded).

 

[status1]

That would be nice but I am not sure if they can control that if the prices are moving so wildly even during trading hours somtimes I see the T+0 line do an out of bounds move for a second than back to normal

I was wondering if ONE or OV have this problem
I guess it depends on where they get the data from

 

[Marcas]

I think this issue is nothing. I mean it's pretty normal, not unreasonable. TOS probably could include solution that Gary proposed but this equals to adding backtesting feature to the platform - what can be complicated to do. TOS do have capability though.

On other hand, what I experienced today, due to the same issue, is that my net liq was fluctuating a lot in early hour. When it showed big negative number I couldn't place any trade. Have to wait for the moment when net liq was positive. Not a biggie but a thing worth to be aware of.

 

[garyw]

That would be nice but I am not sure if they can control that if the prices are moving so wildly even during trading hours somtimes I see the T+0 line do an out of bounds move for a second than back to normal

I was wondering if ONE or OV have this problem
I guess it depends on where they get the data from

While I actually do NOT know what ONE or OV does specifically, the process I use derives the underlying inferred price from the chain, so is not sensitive to this issue.

 

[status1]

derives the underlying inferred price from the chain,

Just curious on how you do that ?
I mean if the option chain is messed up in tos I am not sure how you can derive any accurate prices from that unless you are using some other option chain

 

[garyw]

Just curious on how you do that ?
I mean if the option chain is messed up in tos I am not sure how you can derive any accurate prices from that unless you are using some other option chain

Steve Speer wrote a white paper (chain_bootstrap) a few years ago that discussed techniques for deriving proper option IV/greeks for Index products (with a primary emphasis on properly addressing impact of dividends and interest rates). I implement a number of the techniques he discussed. I am not recommending others do this as it is likely to be considered overkill for most folks.
The real time data needed is BID/ASK for all options in each chain of interest. (TOS provides reliable BID/ASK prices)

The most critical reason for me to use his methods was to properly address the impact of Dividends on the Index option prices. He addresses this by using Call/Put parity for the entire chain of strikes with "usable" BID/ASK prices to extract the proper "U" value to use. He uses the formula U=NDX+PVDIV, where NDX is the spot price, and PVDIV is the present value of the Dividend. TOS still gets this wrong, but TOS is close enough for government work (and most people don't really need the precision).

IMO: The option chain in TOS does contain valid and correct BID and ASK prices for every strike. -- The computed values are wrong when they use incorrect source data, such as wrong dividend value, wrong underlying price for the option and time referenced, etc.

 

[garyw]

I think this issue is nothing. I mean it's pretty normal, not unreasonable. TOS probably could include solution that Gary proposed but this equals to adding backtesting feature to the platform - what can be complicated to do. TOS do have capability though.

On other hand, what I experienced today, due to the same issue, is that my net liq was fluctuating a lot in early hour. When it showed big negative number I couldn't place any trade. Have to wait for the moment when net liq was positive. Not a biggie but a thing worth to be aware of.

That is scary! Was your net liq actually going negative or was that due to "shortcomings" of their code?

 

[D78]

Perhaps one could use thinkback to examine the chains at or near the day's close and then use the option chain (look at prices up or down the chain offset by the difference between the current price and the closing price of the underlying) to derive a better price estimate based on how the underlying has moved after hours?

 

[status1]

(TOS provides reliable BID/ASK prices)

That is mostly true when the markets are open but the discussion here was after hours so that is why I was curious on how you can do that if they are not reliable unles you somehow freeze the data before the close when the bid ask prices are still reliable

 

[garyw]

That is mostly true when the markets are open but the discussion here was after hours so that is why I was curious on how you can do that if they are not reliable unles you somehow freeze the data before the close when the bid ask prices are still reliable

Ah! I see your point! I am only capturing during market hours and only evaluate what I captured, so my references to BID/ASK "always" being correct was not comprehensive, and fell short of what you were seeking. (Sorry for the confusion)

There are a number of "dangers" in trying to evaluate after market hours.

 

[Marcas]

That is scary! Was your net liq actually going negative or was that due to "shortcomings" of their code?

My real net liq was ok.
Due to option price fluctuations and that TOS use prices 'as is' , my net liq was fluctuating a lot. If I remember correctly number was oscillating between +400K and -400K values. When number was negative I couldn't place any trade, had to wait until it become positive, for system to accept order.
As said not a biggie except if net liq was stuck in neg for extended period of time, like right after options cease to trade and you want to place trade with Futures. Then you have to call the desk.

I don't blame code. This issue can be fixed but I afraid a fix can do more harm than help.