International markets traders banter

[shouldaindex]

U.S Markets now only 6% from previous peak.

Thoughts about the direction for the next 3-6 months?

 

[Trembling Hand]

U.S Markets now only 6% from previous peak.

Thoughts about the direction for the next 3-6 months?

"Well, you know this is a bull market!" - Jesse Livermore

 

[CanOz]

U.S Markets now only 6% from previous peak.

Thoughts about the direction for the next 3-6 months?

I'm going to go out on a limb here and call this correction over....

I'm not all skittles n beer, we've got hurdles yet before a larger correction is ruled out....

 

[CanOz]

This looks interesting....

TH, any idea on this?

Deutsche Börse launches Intraday Volatility Forecast for Futures of DAX, EURO STOXX 50 and Euro-Bund

 

[Trembling Hand]

This looks interesting....

TH, any idea on this?

Deutsche Börse launches Intraday Volatility Forecast for Futures of DAX, EURO STOXX 50 and Euro-Bund

Oh great another chart!

I would be sceptical. These new products have a way of being duds. I'm also pretty sceptical of being told by a quant that this addition will help my disc trading :

 

[shouldaindex]

Just doing some research on the S&P500 historical.

Over a rolling 25 year holding period (encapsulating 1926-2015), 95% of the time you returned 7 x your original investment or more.

 

[CanOz]

Range day today or yet a third NR day?

 

[Trembling Hand]

Range day today or yet a third NR day?

Ha! Got your answer yet..

 

[CanOz]

Ha! Got your answer yet..

Indeed, thankfully my trusty CFD platform worked a treat today too, as both CQG and IB had connectivity issues. CQG wouldn't stay connected and IB would give me depth, but no prints....

I managed to open at 10090 and cover at 10023.

 

[sinner]

SPY 100 day correlation between single day return and previous single day return...record breaking would be the correct way to describe it.




mmm...64 slices of American momentum...

 

[CanOz]

I know i bled like a stuck pig after yesty's swing fest...:frown:

 

[Trembling Hand]

SPY 100 day correlation between single day return and previous single day return...record breaking would be the correct way to describe it.

View attachment 64662

Interesting. Could you post a shorter time frame of that chart? (18 months or so)

EDIT: Actually never mind. Just had a better look at it.

 

[sinner]

Interesting. Could you post a shorter time frame of that chart? (18 months or so)

EDIT: Actually never mind. Just had a better look at it.

For those playing at home



and a freebie, FXE

 

[sinner]

A longer study, using S&P500 index data

 

[DeepState]

A longer study, using S&P500 index data

View attachment 64671

Please condition by aggregate return over the rolling window. High correlation probably occurs when there are particularly weak returns. XY plot with Y = 100 day correlation and X = 100 day return might be revealing. Would also correlate with Volatility. I'm quite sure that Howard Brandy (and probably yourself) can explain why this occurs in tractable terms (non-stationary mean during estimation window, with lower underlying mean during stressed times).

 

[sinner]

Please condition by aggregate return over the rolling window. High correlation probably occurs when there are particularly weak returns.

Will check to confirm but don't think so. High correlation in daily returns happens on momentum runs (i.e. if yesterday was up today is up) and low correlation during mean reversion microstructure aka chop.

XY plot with Y = 100 day correlation and X = 100 day return might be revealing. Would also correlate with Volatility.

Will check but don't think so. The future 100 day correlation mostly influenced by past 100 day correlation (high if low, low if high). You can see the most recent decline was quite momentumish, but GFC and tech crash was very mean reversionish. This borne out by equity curves for systems designed to capture short term fluctuations. Think you'll find high vol correlates with extremes in this measure, high or low.

I'm quite sure that Howard Brandy (and probably yourself) can explain why this occurs in tractable terms (non-stationary mean during estimation window, with lower underlying mean during stressed times).

My 2c is that this is influenced by strategy demand or lack thereof. You can see the correlation has always been on the low side, even when high. The microstructure at least since 90s has been very mean reversion supportive. But as demand for strategies to profit off this goes up it forces profitability down (correlation reverses off lows). Exogenous shocks to the market (either up or down) will show up as high correlation while the market drives to new level over consecutive days at which point the demand for short term mean reversion slowly drives it back to lows.

I have a link I want to share but typing this from phone away from desk so this will have to do until I return to my notes

https://quantivity.wordpress.com/2011/07/03/mean-reversion-redux/

EDIT: all applies only to SP500, other markets like FX and commods all behave differently, please don't extrapolate to other markets if you're watching this discussion.

 

[DeepState]

Don't bother, Sinner. On reflection, I think your thoughts are closer to the truth on this. Might look into it a little more given I was off on my assertions. Opportunity to learn something. Thanks.

 

[sinner]

Some messy scatters for DS

100 day rolling returns against 100 day previous day return correlation


100 day historical volatility against 100 day previous day return correlation


These are for SPY not SP500 index data (i.e. only going back to the 90s).

EDIT:

here is the promised link

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1573164

Evaporating Liquidity

The returns of short-term reversal strategies in equity markets can be interpreted as a proxy for the returns from liquidity provision. Analysis of reversal strategies shows that the expected return from liquidity provision is strongly time-varying and highly predictable with the VIX index. Expected returns and conditional Sharpe Ratios increase enormously along with the VIX during times of financial market turmoil, such as the financial crisis 2007-09. Even reversal strategies formed from industry portfolios (which do not yield high returns unconditionally) produce high rates of return and high Sharpe Ratios during times of high VIX. The results point to withdrawal of liquidity supply, and an associated increase in the expected returns from liquidity provision, as a main driver behind the evaporation of liquidity during times of financial market turmoil, consistent with theories of liquidity provision by financially constrained intermediaries.

 

[DeepState]

Some messy scatters for DS

100 day rolling returns against 100 day previous day return correlation
View attachment 64673

100 day historical volatility against 100 day previous day return correlation
View attachment 64674

These are for SPY not SP500 index data (i.e. only going back to the 90s).

EDIT:

here is the promised link

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1573164

Evaporating Liquidity

Thanks. Was looking at stuff over a 20 day horizon this arvo as a result of your posts.

Found as follows:

Returns when auto-correl>0.3 (or thereabouts) are meaningfully lower than when they are <-0.3
- Weaker markets happen with sequences of poor returns.
- Stronger markets tend to occur when prices move more randomly or exhibit day to day over-reaction.
- I was expecting this outcome.

Volatility is clearly inversely related to returns. Market falls tend to be bigger than market rises. I was expecting this outcome.

Yet, Volatility and Auto-Correlation are negatively related. I was not expecting this outcome. In a fit of rushed thinking, I was, instead, thinking about how cross-sectional vol and correlation commove and confusing this with the example you provided which is about auto-correlation and vol.

It would seem that low auto-correlation, then, would be related to high correlation within the market constituents. When macro stuff is moving the entire market around, it is more prone to be over-reacting. This is also what I understand, but this is the first time I have taken this route and seen it this way.

---

Interesting hypothesis in the paper. Turnover jumps a lot during VIX spikes. So it is odd to claim that returns from reversion strategies are attributable to liquidity provision at first glance. I suppose there is money to be made by taking the opposite side of markets whipping around if they mean revert. If there is a lot of panic in whipping around, then the cooler hands will make money. High volume would be the result of high demand for liquidity which is being supplied with a large premium.

Naturally, it is one thing to say we can make money in XYZ market state and predicting that we will remain in that state.

Given volatility persistence, position sizes during these times of high volatility and low auto-correlation would generally be much lower than average, making harvesting dollar profit more difficult than might first appear on a straight conditional sort assuming fixed position sizes in all market regimes.

Provide liquidity when there is blood on the streets....and you know that the blood will stay flowing.

 

[Trembling Hand]

The results point to withdrawal of liquidity supply, and an associated increase in the expected returns from liquidity provision, as a main driver behind the evaporation of liquidity during times of financial market turmoil, consistent with theories of liquidity provision by financially constrained intermediaries.

I think that is a dodgy conclusion. They are saying that because a broken market such as CDO's etc stopped trading there was a equal reduction in equities. You only have to pull up a chart to see that aint right.

There are times when a unit of volume can produce a greater or lesser move/volatility. That is clear for anyone who participates in the market. To claim that a lack of liquidity increased volatility ignores the fact that a greater unit of volume found counter participants. It is not the lack of volume that increase volatility/VIX as they are always associated with increased participation on comparable time frames.

(only read the first few pages)