ETF arbitrage day vs swing trades, eg TQQQ vs SQQQ

Discussion in 'ETFs' started by KCalhoun, Aug 24, 2019.

  1. KCalhoun

    KCalhoun

    I day trade ETF arb pairs like SQQQ/TQQQ, TVIX/SVXY, NUGT/DUST, UWT/DWT, LABD/LABU. See my recent TASC articles. Need to adjust share size based on trading ranges. biggest challenge is figuring out how to scale in/use trailing stops for variety of types of trading days, eg strong trend vs sideways chop vs midday reversals.

    I am 100% convinced that solving it provides a huge advantage, figuring out the details for consistency regardless of mkt condition is my #1 goal. One variant I've traded profitably is to wait til after initial move on strong-trending ORB day, eg TQQQ/SQQQ, then setting buy-stop orders .5 or so above market for each, then scaling into whichever one hits first.

    my biggest failure as trader this last 20 years is extreme risk avoidance, trading small size. its why im still a vendor; i trade small. ETF arbs solves this, because with the safety net of "if one goes up, the other goes down, guaranteed" I have no fear of big stop losses, since gains in one offset losses in the other. solving to get past breakeven in all mkt types is the current challenge. often in chop days oil (UWT/DWT) or gold (NUGT/DUST) works best.

    i haven't posted here in many years (thx Baron for #1 trading forum)... this topic is extremely worthwhile; solving it has the potential to beat the market. solving for swings is likely easier than daytrading. thoughts?
     
    Last edited: Aug 24, 2019
    birdman and murray t turtle like this.
  2. KCalhoun

    KCalhoun

    fwiw here's some notes etfarb1.jpg etfarb2.jpg etfarb3.jpg etfarb4.jpg
     
  3. guru

    guru

    I also suspected this could be a great edge, and I did a lot of work with every leveraged ETF pair - optimizing their trading to no end, and with every possible ratio, mostly focusing on shorting them at the same time, and finding optimal conditions for shorting.
    In the end I did find some alpha but less than trading these ETFs individually, therefore I’ve put the pairs on back burner.
    And I have a suspicion that trading a pair is similar to muting a single leveraged ETF, basically trading less of a single ETF may offer similar risk/return/drawdown as going larger with the pair. On another hand, I might’ve missed something or have bugs that may produce better results once I get back to working on these again. Just wanted to point out that it may not be as worthwhile as it seems.
     
    KCalhoun likes this.
  4. Much better instruments to spread against each other. You didn't say how large your margin equity is. These instruments are not capital efficient!

    1) You want to obtain LONG/SHORT exposures to index like NASDAQ 100.

    You should buy NQ future and sell ES, YM, or RTY futures using same techniques.

    You should be CROSS HEDGING instead of this.

    -THIS IS NOT ARBITRAGE-

    How much money you got?

    --Click on my name and read my thread 'Index Spread Traders'--
     
    KCalhoun likes this.
  5. KCalhoun

    KCalhoun

    good points, makes sense to test ... for intraday I'm testing the idea of 3 decisions/actions each day
    1) morning orb, 9:30-10
    2) check both midday & tighten stops/scale in as needed
    3) end-of-day 3:30-4pm exit w/trailing stops

    correct use of trailing stops & position size add/sub once is the key, plus in vs out mkt day, premkt futs etc

    thx rm, there's also options strategies, plus futs; i like the idea of longNQ vs shortES

    i'm dedicated to solving it (I used to be a statistician/quality guy at ford), surely there's a solution. goal is to keep it simple, consistent and produce greater than breakeven P&L. It would be great to get an active discussion on it.
     
  6. I think that these pairs are too efficient.

    This is very complicated mathematically to pair trade.

    You have the index value, the capitalization of the ETF, then you have the share price, and multiplier on performance....traditional spread trading modelling suggest using implied volatility adjusted notional balancing or "dollar volatility adjusted sizing" of legs.

    My opinion is that there are too many ways that this can go wrong.

    I have a math degree high honors. I love probability theory and mathematical statistics (basically took every class at my school on this).

    My opinion is that timing entry/stop/tp/sl is too emotional style.

    I definitely agree that, working into a position using hedge and careful management and/or scaling will always give you an advantage over participants who time entry and constantly flatten exposure.

    The market doesn't like to give easy entry or tp to flat traders. It likes to find thier stop!

    My opinion is that you should trade NQ with cross hedge YM, ES, or RTY.

    NQ/ES is the most efficient cross hedge because NQ/ES is much riskier to move away from its trend and would require large institutional money.

    NQ/ES rallies under normal bullish market because it represents movement of funds from lower beta stocks into higher beta stocks. :D
     
    Last edited: Aug 24, 2019
  7. KCalhoun

    KCalhoun

    well said; great points! true re "dollar volatility adjusted sizing" is a pita, eg X# shares TQQQ means Y# shares SQQQ since priced differently.

    key is to focus on strongest trending pair of the morning; eg 2day highs; eg it's 10am and SQQQ is up above prior day high, so TQQQ under pd low; Now put in buy stop orders for both & scale into winner ...

    for futs to work the chart pairs would need to be perfectly inversely correlated, like TQQQ vs SQQQ

    swingtrading i've found takes a lot of patience

    youre right re stops in choppy market, so initial entries should be higher
     
  8. I still don't understand what you are doing with entry stop or even how much money you have to trade with.
     
  9. Ok, I'm starting to understand. Here is my opinion. The instruments are not 'perfectly' inversely correlated. Here is why:

    The market groups that are fixing the prices use high frequency trading and special privileges. They have right to buy 'creation units' which is a large block of shares that they can then use to maintain the performance of the ETF. In other words, they are maintaining a 'perfect' hedge competitively. This means that they make money from people trying to use the pair as a hedge and/or cross trade.

    This is complicated, but it would seem that you are increasing transaction costs and requiring multiple instruments only to compete with these groups. They are maintaining market neutral exposure and profitability. While you are just transforming a spread of instruments into a directional QQQ exposure with what I think are unknown weightings (because its so hard to calculate).

    The idea of cross hedging is what you like I believe. I have found that even with my edge I built using programming and markets knowledge, that cross hedging techniques make me a better trader.

    In fact I would say that cross hedging and managing exposure is required to maximize profit from index products. They trade back and forth all day in order to facilitate price discovery. Good Luck!
     
    Last edited: Aug 24, 2019
  10. KCalhoun

    KCalhoun

    good point re cross-hedging; I've tested imperfectly correlated pairs like TVIX / TQQQ and that works on strong but not choppy days... for all-mkt condition they need to be perfect mirror image charts like SQQQ/TQQQ...

    to calculate differential, since priced differently i figure out 'what # of shares would've produced exact breakeven P&L if i bought both at 9:30 and sold both at 4pm", then now i'm testing how to use a midday position sizing/trailing stop approach to produce net profits on the pair.

    like i said in recent TASC article; the inverse correlation is the only guarantee in the market; it's exploitable.

    i honestly believe the solution will beat the market; i'm getting older at 55yo and am big on producing a lot of retirement/travel $ and the solution to this is key. happy to document test results w P&L from trades here in months ahead. im thinking of reaching out to quants to brainstorm, i'm looking fwd to solving it...

    anyone else want to work on it? post away here...
     
    #10     Aug 25, 2019