NexusFi: Find Your Edge


Home Menu

 





Quantitative analysis of speed of moves for ES in most recent vol.


Discussion in Emini and Emicro Index

Updated
      Top Posters
    1. looks_one tpredictor with 7 posts (1 thanks)
    2. looks_two artemiso with 5 posts (4 thanks)
    3. looks_3 iantg with 2 posts (1 thanks)
    4. looks_4 Quick Summary with 1 posts (0 thanks)
    1. trending_up 3,457 views
    2. thumb_up 6 thanks given
    3. group 9 followers
    1. forum 14 posts
    2. attach_file 1 attachments




 
Search this Thread

Quantitative analysis of speed of moves for ES in most recent vol.

  #1 (permalink)
 tpredictor 
North Carolina
 
Experience: Beginner
Platform: NinjaTrader, Tradestation
Trading: es
Posts: 644 since Nov 2011

It feels like the ES is moving insanely fast at times now. Faster then at any time in history due to the HFT algo trading. Can someone run a quantitative analysis of the speed of the futures moves say from 2011 prior and compare to now for the same or similar volatility?

I plan to run my own analysis. But, my feeling is that the market is moving faster then at any time in history. The speed can make it very difficult to trade with discretion.

Started this thread Reply With Quote

Can you help answer these questions
from other members on NexusFi?
New Micros: Ultra 10-Year & Ultra T-Bond -- Live Now
Treasury Notes and Bonds
NT7 Indicator Script Troubleshooting - Camarilla Pivots
NinjaTrader
Better Renko Gaps
The Elite Circle
Deepmoney LLM
Elite Quantitative GenAI/LLM
NexusFi Journal Challenge - April 2024
Feedback and Announcements
 
Best Threads (Most Thanked)
in the last 7 days on NexusFi
Get funded firms 2023/2024 - Any recommendations or word …
61 thanks
Funded Trader platforms
38 thanks
NexusFi site changelog and issues/problem reporting
27 thanks
GFIs1 1 DAX trade per day journal
18 thanks
The Program
18 thanks
  #3 (permalink)
 iantg 
charlotte nc
 
Experience: Advanced
Platform: My Own System
Broker: Optimus
Trading: Emini (ES, YM, NQ, ect.)
Posts: 408 since Jan 2015
Thanks Given: 90
Thanks Received: 1,147


tpredictor,

I have been studying the same thing. Here is what I have noticed:

The major increase in volatility occurred in early February. The size of my market replay files doubled from January to February. There is a much wider daily range, the market order queue has increased and the limit order queue has decreased.

From a macro level here is what I have observed over the last 3 years.

2016: Jan 2016 - Feb 2016: Volatility was high. The market order supply was high enough to drive the price around a lot.

2016: March - December: There was a fair balance between market orders and limit orders.

2017: Volatility dried up because no one wanted to cross the spread. The limit order book was insanely thick at all times, and the market order supply was very low. The price hardly moved.

2018: Jan: Volatility increased significantly compared to all of 2017. I would say that it was similar to Jan-Feb 2016 perhaps.

2018: Feb - Up to now: What I have observed is that the limit order supply has dropped significantly. I think certain classes of algorithms got whacked and operators took these off line. In the past you would see 500-1000 volume on all sides in the DOM. Now it is more like 200-500. I am not sure that this is due to an oversupply of market orders or an under supply of limit orders. It could be a little of both, but my guess is it is likely in part due to certain market making algorithms getting torn up in higher volatility conditions, so the operations just shut them down.

What type of analysis are you looking for exactly? I have some code built that can measure the number of transacted orders per price level, canceled orders per price level, added orders per price level and starting and ending time stamps per price level. This would give you a sense of how much of what is flowing through... and it might give you clues if this more of an oversupply or market orders or an under supply of limit orders behind this.

Ian



tpredictor View Post
It feels like the ES is moving insanely fast at times now. Faster then at any time in history due to the HFT algo trading. Can someone run a quantitative analysis of the speed of the futures moves say from 2011 prior and compare to now for the same or similar volatility?

I plan to run my own analysis. But, my feeling is that the market is moving faster then at any time in history. The speed can make it very difficult to trade with discretion.


In the analytical world there is no such thing as art, there is only the science you know and the science you don't know. Characterizing the science you don't know as "art" is a fools game.
Visit my NexusFi Trade Journal Reply With Quote
  #4 (permalink)
 tpredictor 
North Carolina
 
Experience: Beginner
Platform: NinjaTrader, Tradestation
Trading: es
Posts: 644 since Nov 2011

@iantg This sounds very interesting. As a tape reader, anything regarding the orders might be of interest. However, I'm more thinking about something basic like the amount the market will move in say a 1 second or 30 second interval. That's a first test but the other test might look at how fast the market moves without any pullbacks. I have started to observe many sorts of flash movements. Flash crash. Flash rally. The market behavior itself seems to be still sensible but the speed at which it moves seems to disadvantage any discretionary trader and likely even probably most non co-located automated strategies.

Example might be today from 11:23 to 11:27. The ES made a move a 10 point move with no pull backs visible on a 30 second chart. Reviewing on 30 seconds with ATR. It looks like we had a 4 point move at 11:54 within 30 second. Around 10:10 was also a volatile with 5 point moves occurring within 30 second intervals. We also had another 10 point move without any pull backs visible on a 30 second chart from 9:36 to 9:38.

Started this thread Reply With Quote
  #5 (permalink)
 artemiso 
New York, NY
 
Experience: Beginner
Platform: Vanguard 401k
Broker: Yahoo Finance
Trading: Mutual funds
Posts: 1,152 since Jul 2012
Thanks Given: 784
Thanks Received: 2,685

Times apart consecutive events:

 
Code
ESH8, Mar 1

90 - 5.096 ms
95 - 14.710 ms
99 - 55.691 ms
99.9 - 176.812 ms

ESZ7, Oct 9

90 - 47.253 ms
95 - 121.081 ms
99 - 439.072 ms
99.9 - 1341.504 ms
Considering human reaction time is about 0.1-0.3s, about 99% of the events on ES happen too fast for a person to react. This used to be about 95% on Oct 9, which I randomly picked from last year. There's a numerical difference, but I believe that a colocated system has no material advantage over a human click trader on Mar 1 that it didn't already have on Oct 9. Data wise, the new protocol is actually more efficient. Nov 19, 2014 was about 811 GB raw, whereas Mar 1, 2018, which is huge by recent standards, is only about 657 GB raw.

Reply With Quote
Thanked by:
  #6 (permalink)
 tpredictor 
North Carolina
 
Experience: Beginner
Platform: NinjaTrader, Tradestation
Trading: es
Posts: 644 since Nov 2011

What do you consider an event? Just a market depth change? Right, I'm not referring to that. I'm referring to how fast the market moves when it jumps. I'd like to see a comparison to the recent events to historical events.

I'm curious also if the volume is higher or lower. I suspect that the liquidity provision algos are offering less on the book. But, I'm not sure if they are still there or not, either. If the market is moving too fast for human traders to react too then that's a structural concern, I'd think.

Some questions are:

1. Is the overall volume up or down?
2. Divide the volume by the range. Most technical analysis of volume is all wrong anyway. Because, they don't adjust for the range. Obviously, if the range is more then the volume might be more too because it covers more range. But, the relative volume is more likely to drop. Compare the relative volume for similar binned bars.
3. Is the volume on a price level basis similar to historical times?
4. Look at the range of X second bars. Compare this to when the VIX was at a similar level.
5. Look at the how far/fast the moves without at least an X tick pull back over a given window of time. I think this is the measure I'm most interested in.
6. Count the number of times we test an X high or low as measured from some offset of an average or median. A market that is moving fast will test the high or low multiple times in a shorter period.
7. You could also look at percentage of outside? bars. A bar that takes out the high and low of previous bar. A higher percentage of outside bars would indicate uncertainty but might also suggest more gaming.

Now some suspicions on my part...

First, I think there is more profit opportunity then when the ES wasn't moving. However, the market feels more difficult. I have seen similar behavior in other markets. I believe it is caused by aggressive algorithms that drive the market or paint the tape with aggressive bid/offers. A sign of their activity is I suspect a sawtooth or megaphone pattern with the market making new highs and new lows progressively getting larger and larger. I didn't see that in the ES today but I will look again.

Started this thread Reply With Quote
  #7 (permalink)
 artemiso 
New York, NY
 
Experience: Beginner
Platform: Vanguard 401k
Broker: Yahoo Finance
Trading: Mutual funds
Posts: 1,152 since Jul 2012
Thanks Given: 784
Thanks Received: 2,685

I'm just saying that there's no additional disadvantage over what was already in place in 2017, in response to this:


Quoting 
The market behavior itself seems to be still sensible but the speed at which it moves seems to disadvantage any discretionary trader and likely even probably most non co-located automated strategies.

by quantifying how much 'faster' these events are taking place.


Quoting 
But, my feeling is that the market is moving faster then at any time in history.


Quoting 
I'm curious also if the volume is higher or lower. I suspect that the liquidity provision algos are offering less on the book.

The average spread is actually still fairly tight in the US cash session. It's 1.05 ticks yesterday.

What has changed significantly is that the liquidity at touch is thinner. On a very thin day like Dec 22 (before Christmas close), it's 262, now it's about 54.

Reply With Quote
Thanked by:
  #8 (permalink)
 tpredictor 
North Carolina
 
Experience: Beginner
Platform: NinjaTrader, Tradestation
Trading: es
Posts: 644 since Nov 2011

I guess there is no interest in whether or not possible gaming that occurred at the overnight lows a few nights ago in the ES. Like really, I'm sure that whoever was offering 1,000 lot on the offer after we were down over 45 points wasn't trying to push the market down into stops but was really just wanting to fill their order. Do not look for any possible gaming between 19:32 and 19:36 on what Wednesday morning/Tuesday evening.

But if you do, let me know if you see anything unusual. I personally suspect that either HFT liquidity providers were wrong sided and needed to clear their books and so they drove the market down into stops or perhaps it was just a trader.

Started this thread Reply With Quote
  #9 (permalink)
 iantg 
charlotte nc
 
Experience: Advanced
Platform: My Own System
Broker: Optimus
Trading: Emini (ES, YM, NQ, ect.)
Posts: 408 since Jan 2015
Thanks Given: 90
Thanks Received: 1,147

tpredictor,

If I follow what you are saying, you are looking for some information on what occurred between 7:30 PM and 7:40 PM eastern time Tuesday night. I ran the data for you during the time frame. (Full Depth of Market at the top 10 price levels) .

Here is what I saw:

1. At 7:35:50 PM, At the Ask price = 2683, someone put up 670 to 680 contracts in. At that point the Bid/Ask was 2682.5/2682.75.

2. Around 7:36:16 PM the volume at the price level Ask = 2683 decreased by approximately 670- 680. At that point the Bid/Ask was 2683 /2683.25. It's reasonable to assume this was the same party doing this because it was the same amount.

3. It looks like this party submitted the order initially right after the Bid / Ask spread moved away from that price level where Ask = 2683, and they held their large order in play up until the price was about to get to the Bid Ask spread that would have hit their order, but right before then they canceled the order.

Is this what you are looking for?

Would you characterize this as spoofing?

Let me know if we are talking about the same event or not.

Ian






tpredictor View Post
I guess there is no interest in whether or not possible gaming that occurred at the overnight lows a few nights ago in the ES. Like really, I'm sure that whoever was offering 1,000 lot on the offer after we were down over 45 points wasn't trying to push the market down into stops but was really just wanting to fill their order. Do not look for any possible gaming between 19:32 and 19:36 on what Wednesday morning/Tuesday evening.

But if you do, let me know if you see anything unusual. I personally suspect that either HFT liquidity providers were wrong sided and needed to clear their books and so they drove the market down into stops or perhaps it was just a trader.


In the analytical world there is no such thing as art, there is only the science you know and the science you don't know. Characterizing the science you don't know as "art" is a fools game.
Attached Files
Elite Membership required to download: ES 3-6-18.zip
Visit my NexusFi Trade Journal Reply With Quote
Thanked by:
  #10 (permalink)
 tpredictor 
North Carolina
 
Experience: Beginner
Platform: NinjaTrader, Tradestation
Trading: es
Posts: 644 since Nov 2011


@iantg

Well, I am referring to how a large offer was driving the price. It was actually happening earlier too. It was happening when the market was around 2692 before the support was broken at around 19:04-19:08. There are really only a few explanations.

1. A trader is trying to get filled but cannot because the HFT are running the trade away from them. My problem with this explanation is that in many of these cases the offer appears to be actively driving the market down. It is possible the trader is already in profit and has some sort of chase order to get out of a trade at profit.

2. The orders are some sort of stop-limit order that gets "lit" when price gets too near it. But, at least what it looks like is the market driving. This case would be explained as some sort of liquidation order for a wrong sided trader or one that blew the risk limit. So, that would be an opposite of what I suspect. I have seen similar behavior with other liquidation orders.

3. Sometimes the size is offered on both sides of the book.

4. I am not sure that it is actually spoofing because I have seen in cases where it will get filled. But, I suspect the intention is to signal HFT and other traders to drive the stops. Often the HFT get very aggressive near the lows or highs and then flip the book. It might be both the HFT that is posting the size and driving with the market orders.

5. About the most innocent explanation is a trader wants to get filled on as many as possible so they cover both the book and execute orders and/or they are different traders using stop limits and not a single trader.

At any rate, what I suspect is sorta like getting the HFT to do the bidding so to speak. Once the trader takes a position, they post the large size which they may or may not intend to fill but the purpose is to drive the market into their resting limit buys. A similar explanation is that a large trader gets off side and needs to clear their book. They can do this provided they are trading against weak hands.

But, let me know if you seen spoofing and/or similar behavior around 2692.50-2693 before the support broke. You should find that a large limit order was driving the price down and it was either on the offer or 1 level off only.

Update
@iantg

Yes I have replayed with the help of Ninja 8 and their market depth map. For some reason, the Ninja 7 market replay level 2 data is often wrong and won't show on the depth. We had the same issue with my program and we added a sort of fix/workaround for it. This is only for the market replay data.

Anyway, yes it looks like this trader may have added at 80 and 83 the same trade.

7:35:51 795
7:36:35 564

But, you are right it looks like it stayed on the 83 level though it certainly appears to have drove the market down briefly. If you go back earlier around 7:00 mark you might find one that stayed on the offer and drove the market to break support. Or maybe I imagined that...

Update
Right the big offer came on at 19:35:15 and then 19:35:50 bid showed at 2680. This was right near the lows and took the market below 2682.25 which was the previous low. The ATL was set at 2681 just above the large bid the driver was aiming at. It looks like the driving offers were smaller though, around 85 to 100 contracts only.

Right it looks like the bigger came online when we touched 2682.50 or 1 tick above the prior low. It looks like the order was filled though. At least 569 traded at the 19:40 mark. Perhaps, it was a liquidation order and I was wrong. Or perhaps it was a bug in the software that momentarily lit the order and then it was converted to iceberg or they were able to get filled below the 83 and flipped it. At least 923 traded at the level and the quantity was correct within a few orders of the 700 at least ~200 traded with correct markets on the depth. So my inclination was that it wasn't a spoof order but still might have had a purpose to drive the market lower.

Some other events I marked,
6:03:29 701 2694
6:16:43 237 2693
18:52
19:06-19:07:56 big offers come on 1 tick above the lows...

Update
What is not clear because the N8 isn't updating as fast is whether or not the offers are coming on after we tick the lows or before we tick the lows.
@iantg I think what would answer this if you can see if we break or touch the low before the big offer comes online. If we do then these are probably liquidation or stop limit orders. If the big offer comes online first then it is clearly intended to drive the market down. The one at 19:07:56 is very hard to see what order the events come in. However, I am now leaning to toward the idea these were liquidation orders. It looks like the driving orders are more subtle.

Started this thread Reply With Quote




Last Updated on March 9, 2018


© 2024 NexusFi™, s.a., All Rights Reserved.
Av Ricardo J. Alfaro, Century Tower, Panama City, Panama, Ph: +507 833-9432 (Panama and Intl), +1 888-312-3001 (USA and Canada)
All information is for educational use only and is not investment advice. There is a substantial risk of loss in trading commodity futures, stocks, options and foreign exchange products. Past performance is not indicative of future results.
About Us - Contact Us - Site Rules, Acceptable Use, and Terms and Conditions - Privacy Policy - Downloads - Top
no new posts