kareem40 

and actually add ( yes add) to your losers

That's actually pretty typical in a mean reversion system. The further you are from the "mean" or "normal range" the greater the probability of a reversion, so the bigger the position you want to have on. The problem occurs in those rare times when it doesn't 'revert' as you have you biggest position on as it goes against you!

SMCJB 

That's actually pretty typical in a mean reversion system. The further you are from the "mean" or "normal range" the greater the probability of a reversion, so the bigger the position you want to have on. The problem occurs in those rare times when it doesn't 'revert' as you have you biggest position on as it goes against you!

Hi
Thanks for the post. There is so much negative posts and so called "gurus" that tell you to never to add to a loser. For the most part they are right, I do my very best not to on an outright. But for spread trading and the only thing that have minimized the very big loses is that the fact the I insist on closing all, rain or sunshine, by 3 CST. Testing by hand, I have seen $300-500 daily loss ( based on micro size-max of 5). Not many, you just need to be ready mentally for them when they do happen.
K

Adding to losers is normally a bad idea and I actually wrote a heavily read post [B]Why you should add to winners and never add to losers[/B] a few years back. So why given that post do I think it's both common and okay to add to losers in a mean reversion strategy? Well that analysis assumed a constant probability of the market going up or down. If our hypothesis of a mean reverting price action is correct, this is not the case as the further we get from the mean, the greater the percentage chance that we revert. If I get bored tonite I might write a simulation to try and illustrate this.

In Stock & Commodities May 2002 I found an interesting article about pairs trading.
https://docplayer.net/6892542-Daytrading-stock-pairs.html

Jaap8242 

In Stock & Commodities May 2002 I found an interesting article about pairs trading. https://docplayer.net/6892542-Daytrading-stock-pairs.html

Interesting that they are sophisticated enough look at the correlation and relative volatilities but then trade the spread on a 1:1 basis!

crazymatrix 

I understood everything up until that part and got lost here. Could you give an example or something I could go and read up on further? I was just pointing out that it should start tending towards zero. I would close the trade as soon as it were profitable (so probably long before zero). It's not for no reason. I'm deliberately looking for strategies that are directionally neutral and looking for contracts and situations where I can apply them. I've also been reading up on options, which seem to offer more flexibility in that regard, but options theory looks more complicated. Yes, I have learned about trading various types of spreads with other contracts, e.g. gold/silver, crude oil/natural gas, calendar spreads, etc. The main problem is I only have a $4000 account, so I can't really afford those classic spreads on the major contracts. I mostly trade E-micro gold (MGC), but I keep getting the direction wrong and after a few months I'm mostly break-even. There's no E-micro silver to play off that, so I'm basically looking for two smaller-sized contracts that will work well together. I took this idea from Investopedia -- unfortunately I can't post links yet -- but search for their article "Pairs Trading: Correlation". They have a nice chart there with S&P 500 and Dow as an example and seem to think it's a reasonable pair. Obviously you guys think otherwise. Of course I could trade calendar spreads on MES, but again, as far as I understand, I would have to choose a direction and make either a bullish or bearish trade. One other idea I heard of for smaller accounts was buying a couple of micro contracts, (e.g. MES) then selling an options credit spread on the corresponding major contract (ES) to hedge it and finance it. Have any of you tried something like that? By the way, I realize this is kind of a general thread and not directly related to E-minis. Feel free to move it to a more appropriate subforum.

Yes , if you are trading with someone like Think Or Swim (TOS) with $4000 it is close to impossible to trade spreads. TOS gives you the double Whammy on commissions and margin.
For example for the pair (MNQm21-MESm21) as of today the margin with Interactive Brokers is $741 (one pair) while with TOS it is $2970 . This is typical with TOS.
Interactive Brokers pair ordering portion of the platform works very well. The pair is reduced to a single number with it's own points and ticks. For example at this moment
the (MNQm21-MESm21) is at 1228.40 .

show me a pairs trade and ill show u twice the slippage n twice the commission then ill show you where hft eat your lunch!

wait until you keep adding and they keep diverging! oh boy..double triple the loss! long term capital management couldnt make it work!!!!

but give it a try n report back soon about your big losses ..all the same habits in outright follow u to pairs!!

usually pairs are the last stop in the journey before quitting!!

just my 2 cents!!

SMCJB 

Adding to losers is normally a bad idea and I actually wrote a heavily read post [B]Why you should add to winners and never add to losers[/B] a few years back. So why given that post do I think it's both common and okay to add to losers in a mean reversion strategy? Well that analysis assumed a constant probability of the market going up or down. If our hypothesis of a mean reverting price action is correct, this is not the case as the further we get from the mean, the greater the percentage chance that we revert. If I get bored tonite I might write a simulation to try and illustrate this.

When its mean reversion we say "mean reversion" when its trend continuance, we say "trend continuance" but where is the mean reversion when the trend continuance is working and visa versa?

kevinkdog 

Not to mention one is market value weighted, and the other is price weighted. There are valid pairs to trade, but YM vs ES is likely not exploitable by retail folk.

excellent point.. this is why hft budgets for technology and mathematicians are in teh tens of millions per year!!

datahogg 

Yes , if you are trading with someone like Think Or Swim (TOS) with $4000 it is close to impossible to trade spreads. TOS gives you the double Whammy on commissions and margin.

To this costs possibly add more slippage costs due to the normal slow TOS data delay and then pipeline and local client slow downs with the market hits an impulse.