Anyone scalping the FTSE Futures??

postman said:
Thats quite an interesting statement, let me clarify my position when I say "Prices are always right in this game!"

Imagine the situation where your model says the dax should go down 100 points and you go short, then the dax rallies 200 points and you get stopped out. Try ringing your broker and saying "can I have my money back please 'the price was wrong'!".
Then to rub salt in the wound the dax falls 300 points the next day.

So if the price is always right ....

Your doing a great job - in the short term. I hope you last the long term. :clover:
Click to expand...

That's 47/60 predicted correct. What are you talking about short term long term. The process is perfect....I don't care about the results. Whether I gain money or lose money I know something works or it doesn't if it doesn't I fine tune the model.
Take for example the last hour.....I knew at exactly 2am new york time the ftse will either close at 6173 or 6084.......it has never been off by more than 12 points. So if the ftse is at 6112 and it's 3.30 pm in london........HOW ON GODS GREEN EARTH CAN IT GO 60 POINTS UP IF JESUS HIMSELF DID NOT SHOW UP TO THE LSE?
(in a whisper voice) so I short at anything above 12 points from 6184. Will it close at 6084? Probably not but it sure as hell with not close above 6100.
 
.
 

Attachments

  • jes.jpg
    jes.jpg
    49 KB · Views: 67
joseph1986 said:
That's 47/60 predicted correct. What are you talking about short term long term. The process is perfect....I don't care about the results. Whether I gain money or lose money I know something works or it doesn't if it doesn't I fine tune the model.
Take for example the last hour.....I knew at exactly 2am new york time the ftse will either close at 6173 or 6084.......it has never been off by more than 12 points. So if the ftse is at 6112 and it's 3.30 pm in london........HOW ON GODS GREEN EARTH CAN IT GO 60 POINTS UP IF JESUS HIMSELF DID NOT SHOW UP TO THE LSE?
(in a whisper voice) so I short at anything above 12 points from 6184. Will it close at 6084? Probably not but it sure as hell with not close above 6100.
Click to expand...

What concession do you have for 'outliers' in your Model. ?
 
Mr Sept said:
What concession do you have for 'outliers' in your Model. ?
Click to expand...

To be honest none. I've achieved r values of 1 and the prediction was off. The correlations mainly come from a basket of stocks and some currencies. The problem with outliers is that they usually are associated with news however the current conditions in the currency markets threw my daily predictions off but weekly predictions are still on par. Since I'm doing principle component analysis on the actual stocks vs. ftse the rally this year have confused it....when there is a day it's wrong.....it can't be wrong the next unless central banks meet. The only issue is that the models in the last two weeks have been correct during intraday but closing has been off 5/10 days which hasn't happened. The only other days it was off was the oil news days, belgium terrorist attacks, fomc meeting in march, aussi rate cut and the gold spike. Those days were hard to predict and learned the hard way to not trade news days or unusual spikes. Yesterday intraday was correct but closing was off. The same thing happened today. However sometimes you notice the end of day close is off but otc close is correct like we may see in about 5 hours. The outliers depend on a few factors but outliers can be decreased with a simple hack....increasing the amount of data. So going back all the way to 2002 will be much more accurate than going back to 2008. The only problem is that pattern recognition doesn't seem to understand wierd fluctuations like we are having now.............since it's having difficulty matching the pattern to a said date in the data...case in point the rally from january to april which was only seen in the 1930's.....since the data hasn't been included (Some stocks don't go that far back) we can get things wrong. One thing that is to be noted is that volatility has been increasing steadily since 2014 and this year apparently going to top the last.
If correlations analysis states that 66% correlation with the nodes then we know that if it was wrong yesterday then it must be correct today since 66% correct rate would be every other day of the week. Oddly enough when the correlation is 99% then we know not to trust the model since the indices have been acting irrationally so a rational match would be off. When we have confidence in the model and it states that it will go down it rarely touches the high and vice verse but normally it goes below the low.
 
You must log in or register to reply here.

Similar threads

O
Scalping
Replies
7
Views
4K
roberk
R
O
Futures scalping
Replies
10
Views
4K
neil
neil
D
Pivot Point Scalping on the ES
Replies
7
Views
7K
dancorcal
D
S
Scalping the Dow
2 3 4
Replies
67
Views
19K
Dave1971
D
O
Futures micro trading (scalping)
Replies
1
Views
3K
SuperDriveGuy
S
Top