- ago
Hello Dr Koch,

I'm trying to understand how it is that this strategy determines/predicts that it's time to buy, at the low.

Is it available as a block strategy?
0
1,797
11 Replies

Reply

Bookmark

Sort
- ago
#1
Why not? Buy at limit at 2% below close, sell at market, multiple positions.
0
- ago
#2
Thanks, for that Eugene, but, what are the conditions of the sale? Do I sell on the open at market?

Why can't this strategy be shown as a block strategy if it's so simple, that's my real question I suppose.
0
Cone8
 ( 7.84% )
- ago
#3
No condition. Any position is sold at market the next day.
0
- ago
#4
On the open, right?
0
Cone8
 ( 7.84% )
- ago
#5
At Market orders fill at the open, yes.
0
- ago
#6
pfelix strated a thorough analysis of the OneNight strategy in this discussion thread:
Trying to understand NSF (https://www.wealth-lab.com/Discussion/Trying-to-understand-NSF-11704)

This analysis comes to the conclusion that the OneNight strategy will not show the same performance in real trading as in an EOD backtest.

If you think such unrealistic results should not be advertised in the top 5 or Top 10 lists, then please vote for the following #FeatureRequest:

Add Granular Processing to the Top 5 Rankings (https://www.wealth-lab.com/Discussion/Add-Granular-Processing-to-the-Top-5-Rankings-11770)
0
- ago
#7


Works great. I'm a developer and feel more comfortable writing code-based strategies. But it is super-fast to build stuff using blocks. Going forward I'm going to try to start new strategies using blocks and convert the block strategy to code only if needed.

Your questions about market at open vs limit are very important. Do check out the analysis DrKoch linked to. I learned a lot by going through that exercise. It might answer a few questions and save some time.

How does OneNight determine when to buy? OneNight buys using limit orders. So, the first symbol to reach the limit price will get selected. This will differ from backtesting primarily because backtesting without granular processing randomly selects qualified symbols to take positions in NSF scenarios. This cannot be replicated. You might consider this a form of peeking into the future. Granular processing gets you much close to reality.

Please consider voting for the feature request DrKoch linked to. Keeping the rankings as accurate as possible can only help WL and the entire community.

I also have a less popular feature request. Limit the number of orders that will be generated based on buying power. Brokers are not going to let you place hundreds of limit orders which is another reason OneNight can't be faithfully implemented. We already have the option to use Account Positions in position sizing, this just takes that account information and prevents the unrealistic possibility of placing an unlimited number of active orders. But WL already has probably the best possible workaround in the Quote & Trigger tool.
1
Glitch8
 ( 11.20% )
- ago
#8
I’ll put this idea on the drawing board after recovering from the latest hurricane 🤦🏼‍♂️
0
Cone8
 ( 7.84% )
- ago
#9
Re: Limit the number of orders that will be generated based on buying power.
fwiw, WealthLab actually already does "Pre-Threshold Signal Blocking" for Market order strategies running in Strategy Monitor with "Wait for all Updates". With Market orders we can predict if a Threshold (Preferences > Trading) will be exceeded, and remove the signals (by lowest weight) that would have cause a Threshold to be exceeded.
1
- ago
#10
Very interesting. This sounds like they type of feature that would work for limit orders as well.
0
Cone8
 ( 7.84% )
- ago
#11
Sure it would, but since you don't know which orders are going to hit their target a priori, it would work only to reduce the opportunities. Again, that's a different strategy.
0

Reply

Bookmark

Sort