Historical option prices

[swingstar]

Just wondering if anyone knows where I could get all ASX EOD option prices dating back to 2001?

 

[NettAssets]

Just wondering if anyone knows where I could get all ASX EOD option prices dating back to 2001?

AFAIK these are not available commercially so unless you can find someone who has kept historical records for themselves you will have little luck.

The ASX Database uses the option code as the base index and as this code is reused all the time as soon as an option series expires it is removed from the database.

One other problem is the usefulness of the data when you get it. Unless you screen out all the prices other than actual trades the information can be very misleading.
For example in a longterm series (EG two year expirey) there may be a couple of trades in the first couple of days of its life because it was created on demand. then it may not trade again for 18 months and the last sale price bears little relationship to the actual value and if the underlying has moved substantially then the bid ask will not have been updated.
Much of the price information except for actual trades is theoretical anyway and can be recreated in any option modelling tool.

Good luck in the hunt
John

 

[swingstar]

AFAIK these are not available commercially so unless you can find someone who has kept historical records for themselves you will have little luck.

The ASX Database uses the option code as the base index and as this code is reused all the time as soon as an option series expires it is removed from the database.

Yeah, if they're indexed at all by option code then that'll be a prob.

One other problem is the usefulness of the data when you get it. Unless you screen out all the prices other than actual trades the information can be very misleading.
For example in a longterm series (EG two year expirey) there may be a couple of trades in the first couple of days of its life because it was created on demand. then it may not trade again for 18 months and the last sale price bears little relationship to the actual value and if the underlying has moved substantially then the bid ask will not have been updated.
Much of the price information except for actual trades is theoretical anyway and can be recreated in any option modelling tool.

I was hoping to just get OCH valuation.

Cheers

 

[NettAssets]

I was hoping to just get OCH valuation.

Cheers

You should be able to recreate this with either B/S or Bionomial price models.
Don't know if excell comes with both, I use Lotus and mine has only B/S.
This will make the testing take a lot longer of course than just looking up a precomputed no. but this way you can get aproximate Intra day values which may help a back testing scheme. EOD values will be very misleading on those hammer or long spinning top days. But best of luck in coding it

John

 

[NettAssets]

I do have Hubb optiongear which has backtesting capabilities. On the Aussie market it will only backtest active option series right back to their creation.
I have not used the feature much , Magdoran may have more comments on it, their database is supposed to support backtesting in the next upgrade but I have not found out if it is going to be predated or just carried forward with a new index root.
John

 

[Magdoran]

Just wondering if anyone knows where I could get all ASX EOD option prices dating back to 2001?

Hello swingstar,

It sounds like you are trying to simulate the potential value of options positions and confirm your logic by examining what prices were for specific options in the past to test out your hypothetical forecasts. Is this correct? If so, John’s comments hit the mark…

I fully agree with John’s comments below:

The ASX Database uses the option code as the base index and as this code is reused all the time as soon as an option series expires it is removed from the database.

One other problem is the usefulness of the data when you get it. Unless you screen out all the prices other than actual trades the information can be very misleading.

For example in a longterm series (EG two year expirey) there may be a couple of trades in the first couple of days of its life because it was created on demand. then it may not trade again for 18 months and the last sale price bears little relationship to the actual value and if the underlying has moved substantially then the bid ask will not have been updated.

Much of the price information except for actual trades is theoretical anyway and can be recreated in any option modelling tool.

Part of the problem is that what you really want to know is what the bid and the ask were at a particular price in the underlying at a particular time. This information is not generally available, although it is conceivable that some organisations may keep it for a variety of reasons. For example, I believe that optionsXpress may keep real time data for their US transactions, so they can recreate situations if a complaint or query is made.

If your purpose is to project potential values for positions in the future, I would suggest that an effective way to approach this is along the lines John has suggested by understanding the key “Greeks” involved, and having the capacity to use software to model the potential outcomes to evaluate the potential risk and reward of each position. Currently this is the best “art of the possible” approach available commercially.

Regarding OptionGear, my understanding is that the problem of reusing symbols has been recognised by Hubb, and that they are considering approaches to allowing access to past incarnations of specific options series, but I do not know the timeframe or progress.


Back testing:
As for back testing as a method, my personal position is that it is very difficult to effectively utilise one set of findings that are drawn from past events which had occurred under unique conditions that no longer exist exactly as they did, to forecast future events with much certainty (While not being a fan of Heraclitus, it’s a bit like the idea that you can never step into the same river twice – also it is also aligned with Heisenberg's Uncertainty Principle).

In a wider sense, how you approach this area ultimately boils down to core philosophical beliefs. An alternative to back testing with static one dimensional data, is to utilise modeling for options, and actually looking at charts to develop effective technical analysis skills. Trying to reverse engineer events based on past options performance using a static model, and not factoring the complex effect of key “Greeks” (especially volatility, let alone delta) into the future I suspect is futile – especially if you accept the concepts of chaos theory.

Of course I could be quite wrong here, we may live in a deterministic finite universe, in which case all events could be perfectly modeled, and all events would be predetermined, in which case everything I’m doing is a complete waste of time.

However, if you ask me for my best shot guess right now with what I’ve experienced, I’d have to say the non-finite, non- deterministic view has much more appeal, hence my “belief” if you will embraces Douglas’ view that anything can happen, and every moment in the market is unique, hence I think the conventional attempts at back testing are necessarily flawed and generally have very limited value.

I did say conventional by the way, which means that there are unorthodox approaches which I suspect are much more effective - all who’ve read my comments will know for instance that I’m a subscriber to Soros’ “reflexivity”, and “dynamic disequilibrium” theories as an example. If you subscribe at least in part to these ideas, this opens up a whole raft of radical departures from the current fads.

Anyway, I didn’t mean to go into a philosophical tangent, but when you think about it, it’s all relevant, and when you boil these issues down into their simplest components, the ancient concepts still address these issues.


Regards


Magdoran

P.S. Wayne loves discussions on chaos theory and determinism! Don’t get him started, he’s worse than me (for that matter, don’t get the Duc started either!) Mag

 

[swingstar]

Thanks guys.

I wasn't hoping to test an idea, but rather just see what my results would have been if directionally trading options (just long ATM calls or puts) over the period of time I've been testing (using a light rule-based system - not programmed). I'm more so testing my current understanding of TA and how I could optimise that given different money management schemes and instruments.

To keep things simple I am just using EOD prices, so using EOD OCH valuation I'd still be aware of possible substantial difference between that and MM bid/ask etc., but I don't think the overall result would be too far off the mark.

I have just started Natemberg, so once I have more of an understanding I'll look at using Hoadley's software.

Cheers

 

[lakemac]

sent you a PM about your request.