Excluded Middle

Viewing 10 posts - 1 through 10 (of 10 total)

Author

Posts
10/08/2017 at 8:00 PM #48663 quote
ame
Participant

New
Hi

I’m executing this screener against European Indices (daily), but the problem appears with other lists too.
```
left=highest[20](high[240])   // highest high over one month a year ago
middle=highest[220](high[20]) // highest high during period 1 month to 11 months ago
z=(middle <= left)
//z=(middle >= left)
SCREENER[z](z as "yes")
```
So, I get no results, but even if I comment line 3 and uncomment line 4, I still get no results. This doesn’t make sense to me – has anyone got any ideas what I’m missing?

Thanks
10/08/2017 at 8:57 PM #48666 quote

Despair
Blocked

Master

I don’t know what you want to do but I know that your code does not do what you write in the comments. I think you have to switch the numbers in the [] brackets for your middle variable.

What are the conditions you want to screen for?

10/09/2017 at 1:31 AM #48683 quote

robertogozzi
Moderator

Legend

Line 1 scans for the highest price 240 to 260 bars before.

Line 2 scans for the highest price 20 to 240 bars before and it should be <= that in line 1.

Is that correct?

Despair is right, what do you want to do? Maybe a plain text explanation is clearer than code!

10/09/2017 at 9:25 AM #48711 quote

Nicolas
Keymaster

Legend

Proscreener has a limit of 254 bars of data, try to limit your periods lookback at lines 1 and 2 and it should be ok.

ame thanked this post

10/09/2017 at 10:58 AM #48748 quote

ame
Participant

New

Thanks everyone for all your responses – robertogozzi described exactly what I was trying to do and I think Nicolas has identified the issue – I will try the same on weekly bars to see if the proscreener limitation disappears for the same period.

10/09/2017 at 12:21 PM #48766 quote

Nicolas
Keymaster

Legend

The 254 bars limit exist on all timeframes.

10/09/2017 at 12:43 PM #48771 quote

ame
Participant

New

Yes, I understand, but I’ll be looking over the same date period, so much fewer bars. I’ll post whatever I find. Thank you.

10/09/2017 at 5:42 PM #48796 quote

ame
Participant

New

Thanks very much, that worked exactly as I was hoping. Are prorealtime/proscreener limitations like this documented anywhere?

10/10/2017 at 7:20 AM #48851 quote
ame
Participant

New
Hi again, unfortunately some confusion remains! So, I thought I’d risk another post.

I think there must be something more to it than just the 254 bar limit – see the code below, a much simpler calculation, which uses only 60 bars of daily data. The idea is to find stocks that have hit new highs today but not yesterday.
```
//new high that wasn't a new high yesterday
ha=highest[60](high[1])
hb=highest[60](high[2])
z = (high>ha) and (high[1]<hb)
SCREENER[z](z as "yes")
```
When I run this against “Shares – US Wall St.”, I get many matches, but some of them I believe are in error e.g. “Hawaiian Electric Industries Inc.” – today’s and yesterday’s highs are much lower than the high over the last 60 days so the high>ha isn’t true.
10/12/2017 at 6:48 PM #49193 quote

ame
Participant

New

It looks like the problem has been fixed behind the scenes – running the code today, there are no erroneous matches at all. Thank you to somebody!
Author

Posts