Rules of thumb - trading eod system at intraday price?

[lindq]

In the past I've seen fades on some of these reversion type trades at the end of days, so sometimes it's good to take the mid-day spike, and sometimes it trends on upwards like PRGO did today. As I track it over several # trades I can't tell much difference. Everybody's likely right that I just won't know unless I get some intraday data and test it...

I traded roughly the same strategy for 10 years, and there are a couple pitfalls that you are falling into. (1). EOD data, when trying to buy lows, can be suspect. Daily data can vary depending on the vendor. The point here is, the lows you've seen in backtesting may not have been available to you. If you are serious about this system, you do need to look at intraday data for backtesting to confirm your daily lows, or forward test long enough that you can see if you have a problem between EOD data and reality. (2). What happens on any particular day to any particular equity has NO bearing on the long term success of an EOD system's probabilities. Forcing yourself on the system every moment is not going to help, and can be a big hindrance to your wealth and health.

 

[smili]

I traded roughly the same strategy for 10 years, and there are a couple pitfalls that you are falling into. (1). EOD data, when trying to buy lows, can be suspect. Daily data can vary depending on the vendor. The point here is, the lows you've seen in backtesting may not have been available to you. If you are serious about this system, you do need to look at intraday data for backtesting to confirm your daily lows, or forward test long enough that you can see if you have a problem between EOD data and reality. (2). What happens on any particular day to any particular equity has NO bearing on the long term success of an EOD system's probabilities. Forcing yourself on the system every moment is not going to help, and can be a big hindrance to your wealth and health.

Thanks for the comment. I think I understand what you're saying, but I'm not trying to catch a mid-day low for a buy. I'm placing orders to buy on the open and seem to be getting what I'd expect. (My buy signal is based purely on the market state existing at the end of the previous day.) I only backtest with buy/sell based on open price (for buy) and close price (for sell). I place my buy order then night before to try to be as close to that process as possible.

I'm trying to stay out of the way of the system, but where I'm deviating from backtest is for my selling process. I do out of necessity start checking my signals in the afternoon - about an hour to hour and half before close. That's where I'm introducing a variance to the backtest process. The backtest only sells at the close, but I'm almost always selling prior to the close. It's very possible, like you mention, that I'm hurting results by doing so. Today was a day where prices faded a bit going into the close, but like others have mentioned: one day doesn't make the rule.

From everything I'm hearing here is a) I need to backtest with intraday data if I want to know the tendency for sure, and b) try to implement process as closely to backtest as possible otherwise.

I appreciate the comments from everyone. The market is an interesting puzzle.

 

[lindq]

It is very natural for a new trader to want to micromanage a system, no matter what the timeframe. So don't feel alone in your struggle with this issue.

However...this is one of those areas that you really must master before moving on. If your system sells at the close, then sell at the damn close. Walk away and let it happen. Otherwise you'll continue to insert yourself into your systems and die the death of a thousand cuts.

Now if, after a hundred trades, you do an analysis with intraday data and find that indeed, you would have been better off selling at XXXX time or XXXXX level, then that's a different story. You've approached the subject logically and systematically.

But that's a different story. And that's not now.

So ask yourself. Are you a man who has the courage to trade his system? Or are you a girly-girl little mouse in the corner wetting your panties ?

 

[OddTrader]

Now if, after a hundred trades, you do an analysis with intraday data and find that indeed, you would have been better off selling at XXXX time or XXXXX level, then that's a different story. You've approached the subject logically and systematically.

Agree!