15 replies to this topic

[Rogerdodger]

One observation which I have made on some of the RSI systems, which I list in my blog, is that they can get you in too soon in a sell-off and get you out too soon on a rally.
In the chart below, the problem is real obvious on the Feb. dump.
If you bought just because the RSI 2 hit 10, you would have bought on Feb. 23rd around 1450.
You would have rode it down about 75 points and then sold for a 40 point or so loss when it finally rallied.
But if you waited to buy until the next day's price made a higher high, you wouldn't have bought until May 6th at about 1390.
Could anybody backtest it to buy an oversold RSI, but only when price makes a higher print than the previous candle?
And likewise, only sell overbought when price makes a lower print than the previous candle.
The rules for the three systems are on my blog.

Here is one suggested by vitaminm last night. 2 vs 3 RSI:
http://stockcharts.com/c-sc/sc?s=$SPX&p=D&yr=0&mn=2&dy=0&i=p83184271201&a=103784149&r=4404.png

[borland]

Nice!

You might want to look at the Stochastic RSI which can be used to provide buy/sell signals. Notice how the StochasticRSI put in a higher low after March 12th.

http://stockcharts.com/c-sc/sc?s=$SPX&p=D&yr=0&mn=3&dy=0&i=t41556772887&r=4563.png

[vitaminm]

Rd,

since I don't have an access to intraday stockcharts could you post same chart at 60min or 120min interval to get an early buy/sell signal?

One observation which I have made on some of the RSI systems, which I list in my blog, is that they can get you in too soon in a sell-off and get you out too soon on a rally.
In the chart below, the problem is real obvious on the Feb. dump.
If you bought just because the RSI 2 hit 10, you would have bought on Feb. 23rd around 1450.
You would have rode it down about 75 points and then sold for a 40 point or so loss when it finally rallied.
But if you waited to buy until the next day's price made a higher high, you wouldn't have bought until May 6th at about 1390.
Could anybody backtest it to buy an oversold RSI, but only when price makes a higher print than the previous candle?
And likewise, only sell overbought when price makes a lower print than the previous candle.
The rules for the three systems are on my blog.

Here is one suggested by vitaminm last night. 2 vs 3 RSI:
http://stockcharts.com/c-sc/sc?s=$SPX&p=D&yr=0&mn=2&dy=0&i=p83184271201&a=103784149&r=4404.png

[Rogerdodger]

I don't think hourly charts can be linked directly. I can see it but can you?
If not try this link: http://stockcharts.c...p...1459&r=5244
http://stockcharts.c...1459&r=7561.png
I was doodling with those red lines marking the low of each day to be the sell stop for the next day.
Although today did not hit yesterday's stop, I sold because of the BB's and the general "overboughtfulness".

If that link doesn't work let me know & I'll upload it on imageshack.
Roger

Edited by Rogerdodger, 17 April 2007 - 10:14 PM.

[borland]

http://bigcharts.mar...&mocktick=1.png

[vitaminm]

I don't think hourly charts can be linked directly.
Try this link: http://stockcharts.c...p...1459&r=5244

If that doesn't work let me know & I'll upload it on imageshack.
Roger

It does work.
SSO
I see Rsi2<Rsi3 on 60min chart while on daily Rsi2>Rsi3 May lead to short term pullback.
also Its up for 4days in row!!
Thanks.

Edited by vitaminm, 17 April 2007 - 10:22 PM.

[crestdorf]

As an aged veteran of trading system building and the indepth study of backtesting methodology, I would urge you to ONLY test using forward data. There are several testing systems out there which will allow you to do so. DO SO! I have created more grails than you can imagine. I have raised OPM funds using "backtests" and have found out the hard way what you get by relying on that type of testing.

[Rogerdodger]

I paid good money to attend a sales pitch for a testing system at the Vegas Traders Expo. I figured out that I can't figure it out! In the back of my mind I kept thinking: why is this guy here pushing this miracle system instead of lying on his own private beach in Bora Bora? So back testing has it's value but as you know, It's no holy grail. As I am writing this, I have a friend applying for a position in sales of options training and systems. The lady can't make any money doing it but she's going to teach others.

Edited by Rogerdodger, 17 April 2007 - 11:13 PM.

[vitaminm]

As an aged veteran of trading system building and the indepth study of backtesting methodology, I would urge you to ONLY test using forward data. There are several testing systems out there which will allow you to do so. DO SO! I have created more grails than you can imagine. I have raised OPM funds using "backtests" and have found out the hard way what you get by relying on that type of testing.

One may check his/her trading signal here.

http://www.howthemar...s.com/index.htm

[redfoliage2]

One observation which I have made on some of the RSI systems, which I list in my blog, is that they can get you in too soon in a sell-off and get you out too soon on a rally.
In the chart below, the problem is real obvious on the Feb. dump.
If you bought just because the RSI 2 hit 10, you would have bought on Feb. 23rd around 1450.
You would have rode it down about 75 points and then sold for a 40 point or so loss when it finally rallied.
But if you waited to buy until the next day's price made a higher high, you wouldn't have bought until May 6th at about 1390.
Could anybody backtest it to buy an oversold RSI, but only when price makes a higher print than the previous candle?
And likewise, only sell overbought when price makes a lower print than the previous candle.
The rules for the three systems are on my blog.

Here is one suggested by vitaminm last night. 2 vs 3 RSI:
http://stockcharts.com/c-sc/sc?s=$SPX&p=D&yr=0&mn=2&dy=0&i=p83184271201&a=103784149&r=4404.png

Roger: Do you must use SharpChart2 to do the overlay of the indicators on each other? Which is a member-only service from StockCharts.com? Thanks.

Edited by redfoliage2, 18 April 2007 - 04:48 AM.