T-day expectancy

[opt789]

Quote from Pekelo:

That's what I have been saying, see the opening post:

A nebulous concept at best.

Over the past year, 48.8% of the trading days were up while 50.0% were up for T days.
So I guess you can say that there is a slight bias, 1%, just based on the close.

 

[Pekelo]

Quote from opt789:


So I guess you can say that there is a slight bias,

...and that is exactly what my research shows.

The question is: Is there any way how we can exploit this information?

Well, I also started a thread about cup and handle patterns in this subforum. Out of the 5 days when we had this pattern this year, 3 of them were on Tuesdays.

See where I am going with this? If there is a decent selloff on a T-day in the first 2-3 hours, based on this theory we can expect with a high probability a rally or at least a decent bounce most of the time...

 

[tomahawk]

I generally like your ideas Pek, but one flaw I see with your current test is that you're doing it on data from a period where for 4-1/2 of the last 5-6 months the market rallied 50%! I would think this might skew the results toward the positive.

 

[Pekelo]

Quote from tomahawk:

you're doing it on data from a period where for 4-1/2 of the last 5-6 months the market rallied 50%! I would think this might skew the results toward the positive.

I thought about that. The rest of my charts are on a different harddrive, so until I hook that up, I don't have access to them. But I will also check a bearish period and compare the two.

In the long run I still expect a slightly positive outcome for T-days...

One other reason I am doing this manually and slowly, because I also checked for the rest of the week to see if there is a statistically signifficant preference for rally or selloff on a particular day...

Here is the Summary of the last 5 months:

Tuesdays: 7 up, 6 down, 9 variable

Thursdays: 11 up, 6 down, 4 variable

So Tuesdays have no statistical signifficance, but Thursdays (specially if tomorrow is an up day) are more likely to go up then down with a high probability...

 

[Pekelo]

Quote from Pekelo:

Here is the Summary of the last 5 months:

Thursdays: 11 up, 6 down, 4 variable

but Thursdays (specially if tomorrow is an up day) are more likely to go up then down with a high probability...

We have a huge up day so far, so the final tally for

Thursdays: 12 up, 6 down, 4 variable

Quote from Vespasian:

I had a lot of respect for you till I read this post.

You were sayin'?

I think you should have waited until the research's conclusion is drawn and the theory is either busted or proved.

Me thinks you talked too soon....

 

[ivanbaj]

Don't Count Your Chickens (Until They've Hatched)

 

[Pekelo]

It really doesn't matter how the day ends, had you been long from the morning or last night, you would be in profit and there was a signifficant upmove already.