A first look at auto-trading Tim Sykes using Covestor Investment Management

About a month ago I signed up for Covestor Investment Management (CV.IM). I put $10,000 into a new account at Interactive Brokers controlled by them and then signed up for that $10,000 to be auto-traded following Timothy Sykes. I mainly did this out of curiosity to see how well the trade replication worked. Please note that only Sykes’ trades in one of his two Interactive Brokers accounts are replicated by those following his strategy on CV.IM (trades in his other IB account are replicated with the CV.IM Tim Sykes Aggressive strategy, while trades in his SogoTrade and ThinkorSwim accounts are not replicated anywhere).

My biggest problem with CV.IM is that it subdivides orders and then sends those orders at two minute intervals, guaranteeing that the last shares filled will be filled long after the first shares were filled [Edit 12/10/09 see comment #7 below for an explanation from Covestor]. While this may make sense to minimize market impact for longer-term traders, it is absurd to have an order-replication algorithm that does such a thing when trying to track a day-trader of volatile stocks. Even ignoring the problems with subdividing the orders, out of 7 total trades (three round-trips), 2 had initial fills over 15 minutes after Sykes’ order was filled (the initial short on DRAM and the buy on LZB). Because of this risk of delayed replication time I must emphatically recommend against signing up with CV.IM to follow Tim Sykes or any other short term trader.

Below are my round-trip trades so far and the comparisons to where Tim Sykes made the trade and when I received an alert from his TimAlerts trade alert service. First I compare my profit margin per trade with Sykes’ profit margins:

IMGG —  Me: +14.3%
IMGG — Tim: +13.1%
Difference: +1.2% in my favor

DRAM —  Me: (8.3%)
DRAM — Tim: +5.6% (approximate)
Difference: (13.9%) to my detriment

LZB —  Me: (4.4%)
LZB — Tim: (2.0%) (approximate)
Difference: (2.4%) to my detriment

From these three trades I have lost a net of $30.78, including commissions and short stock borrow fees.

The first trade, IMGG, was replicated quickly:

My CV.IM account short sold 571 IMGG on 11/17/09 at  $1.645 at 9:34:40
My CV.IM account covered 571 IMGG on 11/18/09 at $1.41 at 9:36:20

Tim Sykes short sold IMGG on 11/17/09 at $1.64 and I received the TimAlert email at 9:39
He covered IMGG on 11/18/09 at $1.425 and I received the TimAlert email at 9:38

For IMGG, everything worked well and both sides of the trade were replicated quickly and almost exactly at the same price Tim Sykes trades were at.

The second trade, DRAM, was not replicated quickly:

My CV.IM account short sold 131 shares of DRAM at $4.28 at 12:03:07.
My CV.IM account short sold 146 more shares of DRAM at an average price of $4.53 between 13:50:19 and 13:58:34. This was submitted as four separate orders spaced 2 minutes apart on average.
My CV.IM account covered all 277 shares of DRAM at an average price of $4.746 between 9:42:18 and 9:48:26. This was submitted as four separate orders spaced 2 minutes apart on average.

Tim Sykes short sold DRAM on 11/23/09 at $4.68 and I received the TimAlert at 11:46.
He short sold more DRAM on 11/23/09 at $4.59 and I received a TimAlert (text only) circa 14:25.
Tim covered all his shares of DRAM on 11/24/09 at $4.40ish and I received a TimAlert email at 9:47.

For DRAM, the first short trade was over 17 minutes after the TimAlert (so probably just about 20 minutes after Sykes actually made the trade). It was also at a price over 8% lower than Sykes’ trade. Because I do not have adequate records of when I received the second short alert, I cannot say how close the first trade at CV.IM was to it. However, there was still 8 minutes between first and last fill on the subdivided short orders. The cover orders began before I received the TimAlert, but the fills on the buy to cover orders were still over 7% higher than Sykes’ cover.

The third trade, LZB, was not replicated quickly:

My CV.IM account bought 123 shares of LZB on 12/8/09 at an average price of $10.125 between 15:24:18 and 15:40:28. This was submitted as five separate orders spaced 3 minutes apart on average.
MY CV.IM account sold those 123 shares of LZB on 12/9/09 at an average price of $9.761 between 9:52:15 and 10:00:16. This was submitted as five separate orders spaced 2 minutes apart on average.

Tim Sykes bought LZB on 12/8/09 at $10.00 and I received the TimAlert email at 15:06.
He sold LZB on 12/9/09 at $9.80ish and I received the TimAlert email at 9:55.

For LZB, even the initial buy order from CV.IM came 25 minutes after Tim Sykes bought and 18 minutes after I received the TimAlert. The fills on the sell were better, with the initial fills beating the TimAlert, but because of the spacing of the subdivided sell orders, many of the shares were not sold until after the TimAlert was received.

Here are the details from each of my trades copied from my Interactive Brokers activity statement:

Symbol Date/Time Exchange Quantity T. Price C. Price T. Amount Comm/Tax Basis Realized P/L MTM P/L Code
DRAM 2009-11-23, 12:03:07 -131 4.2800 4.4800 -560.68 -1.00 -559.68 0.00 -26.20 O
DRAM 2009-11-23, 13:50:19 -57 4.5900 4.4800 -261.63 -1.00 -260.63 0.00 6.27 O
DRAM 2009-11-23, 13:52:35 -39 4.5900 4.4800 -179.01 -1.00 -178.01 0.00 4.29 O
DRAM 2009-11-23, 13:56:59 -28 4.4300 4.4800 -124.04 -1.00 -123.04 0.00 -1.40 O
DRAM 2009-11-23, 13:58:34 -22 4.4000 4.4800 -96.80 -1.00 -95.80 0.00 -1.76 O
DRAM 2009-11-24, 09:42:18 57 4.5750 5.3900 260.78 -1.00 243.52 -18.25 46.46 ML;C
DRAM 2009-11-24, 09:44:52 11 4.6900 5.3900 51.59 -1.00 47.00 -5.59 7.70 ML;C
DRAM 2009-11-24, 09:46:23 5 4.6400 5.3900 23.20 -1.00 21.36 -2.84 3.75 ML;C
DRAM 2009-11-24, 09:48:26 204 4.799608 5.3900 979.12 -1.02 905.28 -74.86 120.44 P;ML;C
Total DRAM 0 92.53 -9.02 0.00 -101.54 159.54
IMGG 2009-11-17, 09:34:40 -571 1.6450 1.4850 -939.30 -2.86 -936.44 0.00 91.36 O
IMGG 2009-11-18, 09:36:20 571 1.4100 1.2100 805.11 -2.86 936.44 128.47 -114.20 ML;C
Total IMGG 0 -134.18 -5.72 0.00 128.47 -22.84
LZB 2009-12-08, 15:24:18 22 10.1400 10.1100 223.08 -1.00 224.08 0.00 -0.66 O
LZB 2009-12-08, 15:26:17 22 10.1300 10.1100 222.86 -1.00 223.86 0.00 -0.44 O
LZB 2009-12-08, 15:28:14 17 10.1300 10.1100 172.21 -1.00 173.21 0.00 -0.34 O
LZB 2009-12-08, 15:36:14 11 10.1000 10.1100 111.10 -1.00 112.10 0.00 0.11 O
LZB 2009-12-08, 15:40:28 51 10.1200 10.1100 516.12 -1.00 517.12 0.00 -0.51 O
LZB 2009-12-09, 09:52:15 -5 9.8200 9.8700 -49.10 -1.00 -50.95 -2.85 -0.25 ML;C
LZB 2009-12-09, 09:54:16 -5 9.8200 9.8700 -49.10 -1.00 -50.95 -2.85 -0.25 ML;C
LZB 2009-12-09, 09:56:17 -5 9.7900 9.8700 -48.95 -1.00 -50.95 -3.00 -0.40 ML;C
LZB 2009-12-09, 09:58:19 -97 9.7600 9.8700 -946.72 -1.00 -985.98 -40.26 -10.67 ML;C
LZB 2009-12-09, 10:00:16 -11 9.7000 9.8700 -106.70 -1.00 -111.54 -5.84 -1.87 ML;C
Total LZB 0 44.80 -10.00 0.00 -54.80 -15.28
Total in USD 3.14 -24.74 0.00 -27.87 121.42

Disclosure: No positions in any stocks mentioned in this post. I currently have $5,200 subscribed to follow Tim Sykes at CV.IM. I receive a commission for every DVD or TimAlerts subscription bought through my affiliate link to Sykes’ web store. To see more details on my relationship with Sykes, please see my disclosures. I have a disclosure policy and you can find all my disclaimers there as well; those disclosures & disclaimers are incorporated by reference into this post.

12 thoughts on “A first look at auto-trading Tim Sykes using Covestor Investment Management”

  1. I knew the covestor autotrade lagged, but that is worse than i thought it was. I have always thought it was a bad idea. My opinion is that if you don’t have time to be watching your stocks then you should not be trading. Personally when I get busy at work I just ignore the markets all together. I usually just make 1-2 trades a week since we have been pretty busy at work.

  2. I think CV.IM might be a good idea for following long term trades, I wouldn’t do it for day trading and specially not for following Tim’s trades.

    For one reason or another, the entry & exit prices of his trades are very difficult to match and a few minutes of difference might eat all possible profits.

    As a person I can decide whether or not to go on with a trade, but I think covestor’s system will always follow him even if the R/R ratio has totally changed.

  3. Hi Michael

    I am one of the founders of Covestor and wanted to respond personally. We don’t sub-divide orders, our real time feed from the model manager’s brokerage account provides us with the individual fills for a single manager trade (which may not all be executed at the same time). Client orders are generated once these fills have passed through a series of filters.

    You are absolutely right to say this can be optimized and our new algorithm will amongst other things aggregate model manager fills when executed at the same time, reducing the number of client orders. This should be implemented within the next few weeks.

    Of course when a high frequency trader is trading relatively illiquid stocks the broker is not always able to fill orders immediately which can result in delays. But overall drift is small.

    I believe our new algorithm will address your main concerns and I hope you will give it a try in a few weeks when it’s in place to see for yourself.

    We appreciate the feedback as we are continually striving to improve our processes across all trading strategies so if you have any further questions or concerns I’d delighted to address them if I can.


  4. Ive been a subscriber with the CV autotrade with tim sykes and can confirm that my account performance has been no where near the posted returns of the model.

    Im up less than 2% over the last 3 months with his autotrade system. His posted returns are

    Monthly Performance (%)
    Nov 9.68
    Oct 2.35
    Sep 3.99
    Aug 4.90

  5. I know the original post was from 2009 but wondering if you’ve done any analysis on CV’s supposed new algorithm? I’m thinking about opening a CV account to mirror relatively short-term traders but so far it doesn’t sound very promising.


Leave a Reply

Your email address will not be published. Required fields are marked *

Please complete the formula below to prove that you are human * Time limit is exhausted. Please reload CAPTCHA.

This site uses Akismet to reduce spam. Learn how your comment data is processed.