That's a tough one. I know from experience, it's a lot of small wins and the occasional huge loss. We had a small team of 6 guys about 15 years ago, running an algo based on enveloping. It was scary as shit. That's probably why it worked. You have a thousand standing orders out there to fade the big order. We all did very well. (none of us were long during the flash crash. We had all stopped using the algo by then.) We just had to quit. And then our developer lost around 35k in 10 seconds because of a buyout. I just don't know if would recommend enveloping in a meme stock world.
What would be a good filter for a Mean-Reverting system?
- Thread starter Oadmani
- Start date
That's a tough one. I know from experience, it's a lot of small wins and the occasional huge loss. We had a small team of 6 guys about 15 years ago, running an algo based on enveloping. It was scary as shit. That's probably why it worked. You have a thousand standing orders out there to fade the big order. We all did very well. (none of us were long during the flash crash. We had all stopped using the algo by then.) We just had to quit. And then our developer lost around 35k in 10 seconds because of a buyout. I just don't know if would recommend enveloping in a meme stock world.
%%
Sounds like a winner;
except ''desired position'' could easy be a polar bear trap.
Also i'Ve studied polar bears for decades, common pattern\ the younger ones break in houses+ get shot like a red dead duck.
BUT was surprized in 2021 to see a normal slow polar bear @ a seal hole;
instead of waiting until the seal gets in paw range/ he jumped fast in that seal hole amazing catch\ seal meal polar bear.
,.
Actually, polar bear population is doing fine;
they dont believe in global, socialist scam warming \LOL
%%
WOW;
you were trading like a Wolverine market maker. Meaner than bear + QQQuicker/LOL
FLASH crash looked like a fake quotes to me, in 2020 hindsight but frankly i'm never as quick as a wolverine.
BUY monday helps in the stock market even though that can easy be 100% or 66.6%wrong; + 52 mondays in the years, could be right 70% or 100% wrong/losses in a month or 2.........
You are not providing enough information.
Are you trading both long and short?
Was the losing streak during the 2008 melt down just a function of your system not adjusting to the increased volatility? For most mean reversion traders those were excellent times.
That's not enveloping. There is nothing you could ever draw on a chart to describe what we did. Bollinger bands is just more nonsense a child could do with a paint brush.
We can't short stocks where I trade. Therefore, it is a long-only strategy.
Yes, my method doesn't adjust for increased volatility. Any thoughts how I can do this?
A simple filter could be from a trend indicator, such as using the slope direction of the 50EMA. I only buy stocks as trades when both they and the index have 50EMA's sloping upwards.
I also avoid trading forex when the majority of recent daily price moves across the 28 main pairs are contrary to the sequence of 20 and 50EMA's (in an uptrend, the 20 would be above the 50, and in a downtrend below).
Knowing when not to trade, as your question implies, is vital - more so than the simplistic entry patterns we all started out with.
I also avoid trading forex when the majority of recent daily price moves across the 28 main pairs are contrary to the sequence of 20 and 50EMA's (in an uptrend, the 20 would be above the 50, and in a downtrend below).
Knowing when not to trade, as your question implies, is vital - more so than the simplistic entry patterns we all started out with.
If you let me know your risk/reward ratio, it would help make commentary on filters.
I've done a lot of back-testing in my time, trying to improve performance of systems with the application of filters. They are not really the answer. It's very, very, very difficult to adapt a system with a filter that results in a long term edge that beats the implied odds of the win rate.
I only ask because your win rate of 67% is almost nearly bang on the implied probability of 0.5:1, i.e. risking 1 for a return of 0.5.
If that's the case, applying a filter will by nature of the application of the filter, get you some more gains in some cases, but also make you miss out on some trades you otherwise would have taken, that turn out to the be winners, but are now excluded as wins from your result set.
That will culminate in your win percentage remaining around the same as the implied probability of your risk reward ratio. i.e. no long term edge.
TDUK