RT 20DMF spikes at close?

[Harry]

Pascal,

I have a question regarding the dramatic spikes EOD in the RT 20DMF. I would like to take advantage and trade using RT data, but will not until I understand what's going on under the hood. Without asking you to disclose proprietary information, can you please generally shed some light on the intraday calculation? From what I gathered reading your book, EV is calculated using a complete days data. I can guess at two ways to calculate RT values, either using the previous day/average days EV or treating the current days data as it rolls in as EOD data.

Is the spike an artifact of the calculation method (implying earlier during the day signal is not as reliable) or a true reflection of the market at the close? It's difficult for me to trade knowing that there are major spikes at the end of the day that should stop out a trade but, without actively monitoring from 3:55-3:59, I may be trapped in a trade overnight. Can you simply share your insight as to why we typically see EOD spikes?

A second question is a general one. Since the 20DMF cover 97+ sectors including defensive areas (bonds, etc.), where does the money go when 20DMF goes negative, just cash?

Harry

[Pascal]

Pascal,

I have a question regarding the dramatic spikes EOD in the RT 20DMF. I would like to take advantage and trade using RT data, but will not until I understand what's going on under the hood. Without asking you to disclose proprietary information, can you please generally shed some light on the intraday calculation? From what I gathered reading your book, EV is calculated using a complete days data. I can guess at two ways to calculate RT values, either using the previous day/average days EV or treating the current days data as it rolls in as EOD data.

Is the spike an artifact of the calculation method (implying earlier during the day signal is not as reliable) or a true reflection of the market at the close? It's difficult for me to trade knowing that there are major spikes at the end of the day that should stop out a trade but, without actively monitoring from 3:55-3:59, I may be trapped in a trade overnight. Can you simply share your insight as to why we typically see EOD spikes?

A second question is a general one. Since the 20DMF cover 97+ sectors including defensive areas (bonds, etc.), where does the money go when 20DMF goes negative, just cash?

Harry

I have no specific explanation for the EOD spike.
These could be due to:
- Sell/Buy "at the close" orders. These are thus executed together in the last trading minute. I suspect that many ETFs must execute trades at the close.
- HFT's are closing their daily positions
- Data discrepancies

It would be easy for us to filter these last minute orders out, but would it be wise?
These spikes have been there in the past too. Simply we could not see them.

Regarding your second question, the 97 sectors only include equities. No bonds, no futures, no currencies.
I do not know where the money is invested in when it flows out of equities.


Pascal

[Harry]

Thank you for the explanation.