Scribd Logo
Upload

Welcome to Scribd!

  • 229 views

    Worldquant Report

    Uploaded by

    themaker1602
    The document provides examples of alpha signals that can be used on the WorldQuant Brain trading platform. It describes various trading concepts like turnover, decay, and truncation. It then gives examples of alpha signals for US and Chinese markets, including signals based on price, volume, sentiment, and volatility.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    Worldquant Report

    Uploaded by

    themaker1602
    229 views7 pages
    The document provides examples of alpha signals that can be used on the WorldQuant Brain trading platform. It describes various trading concepts like turnover, decay, and truncation. It then gives examples of alpha signals for US and Chinese markets, including signals based on price, volume, sentiment, and volatility.

    Original Title

    Worldquant_Report

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    The document provides examples of alpha signals that can be used on the WorldQuant Brain trading platform. It describes various trading concepts like turnover, decay, and truncation. It then gives examples of alpha signals for US and Chinese markets, including signals based on price, volume, sentiment, and volatility.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    229 views7 pages

    Worldquant Report

    Uploaded by

    themaker1602
    The document provides examples of alpha signals that can be used on the WorldQuant Brain trading platform. It describes various trading concepts like turnover, decay, and truncation. It then gives examples of alpha signals for US and Chinese markets, including signals based on price, volume, sentiment, and volatility.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    You are on page 1of 7

    Alpha Trading with the

    WorldQuant Brain Platfrom

    Alexis D. Plascencia

    E-mail: [email protected]

    Abstract: We provide a guide to use the WorldQuant Brain platform. We overview some
    trading concepts and give some explicit examples of alpha (α) signals.
    Contents

    1 Introduction 1

    2 Trading 1
    2.1 Turnover 1
    2.2 Decay 1
    2.3 Truncation 2
    2.4 Datafields 2
    2.5 Operators 2

    3 Alphas 2
    3.1 American market 2
    3.2 Chinese market 5

    1 Introduction

    WorldQuant Brain [1] is an online simulation platform to build alpha (α) signals and test
    them using market data. However, it is hard to find resources and documentation on the
    use of this platform. In this report we overview and provide some examples of alpha signals.

    2 Trading

    2.1 Turnover
    Reducing turnover. In order to reduce turnover we can add a condition on the trading.
    Instead of trading at each market open, it will trade only when event is True:

    event = volume>adv20;
    alpha = news_cap;
    trade_when(event, alpha, -1)

    2.2 Decay
    Decay: Averages the alpha in a time window. For example if the alpha=1 from Monday
    to Thursday and then in Friday we have alpha=10, there will be suddenly a lot of trading.
    But if decay=5, then this will take the alpha average over the last 5 days so instead of 10
    it iwll do 10+1+1+1+1/5=2.8 which be less trading.

    Increasing the Decay helps to reduce the Turnover.

    –1–
    2.3 Truncation
    This is the maximum weight for a given stock. It can limit the maximum loss; reducing
    this parameter helps with the concentration.

    The following instruction will tell you will tell you what’s the maximum weight assigned:

    max(alpha) > 1 : 1 ? 0

    2.4 Datafields
    There are three types of datafields in the BRAIN platform:

    • Matrix: Market data is in a matrix. X: stocks Y: time. Can be used to create alphas
    directly based on datafields descriptions.

    • Vector: Data stored accross three axis. X: stocks Y: time Z: time (intraday).

    • Group: Stock classification based datafields. Can be used to improve specific alpha
    performance by neutralizing the alpha by the datafields.

    2.5 Operators
    Different operators:

    vector_neut(a,b)

    This operator calculates a new vector a2 such that a2 is orthogonal to b. It can be used to
    neutralize an alpha vector over some particular risk.

    3 Alphas

    3.1 American market


    1. Trade based on volume and price difference. If the volume is larger than the 20-day
    average, then trade based on the price difference of the last 5 days.

    event = volume>adv20;
    alpha = (-ts_delta(close,5));
    trade_when(event,alpha,-1)

    Region: USA
    Universe: Top3000
    Delay: 1
    Neutralization: Subindustry
    Decay: 2

    –2–
    Truncation: 0.01
    Pasteurization: Off
    Unit Handling: Verify
    NaN Handling: Off

    2. A simple Price Reversion alpha

    SMA_30 = ts_mean(close,30);
    rank(SMA_30 - close)

    Region: USA
    Universe: Top3000
    Delay: 1
    Neutralization: Subindustry
    Decay: 4
    Truncation: 0.08
    Pasteurization: Off
    Unit Handling: Verify
    NaN Handling: Off

    3. Price weighted average. vwap is the daily volume weighted average price. The fol-
    lowing alpha uses vwap to implement Price Reversion:

    (vwap-close)/vwap

    Region: USA
    Universe: Top3000
    Delay: 1
    Neutralization: Subindustry
    Decay: 15
    Truncation: 0.08
    Pasteurization: On
    Unit Handling: Verify
    NaN Handling: Off

    In Fig. 1 we present the results for this alpha signal.

    4. More on price reversion but with a condition on the volume

    –3–
    Figure 1: Results for the price weighted average alpha.

    twenty_day_SMA = ts_mean(close, 20);


    trade_when(volume>=ts_sum(volume,5)/7.5,
    rank(twenty_day_SMA-close),-1)

    Region: USA
    Universe: Top3000
    Delay: 1
    Neutralization: Subindustry
    Decay: 4
    Truncation: 0.08
    Pasteurization: On
    Unit Handling: Verify
    NaN Handling: Off

    5. Implied volatility.

    alpha = ts_decay_linear(ts_delta(implied_volatility_call_60, 25)


    > 0, 20);

    Region: USA
    Universe: Top3000
    Delay: 1
    Neutralization: Sector
    Decay: 4
    Truncation: 0.08
    Pasteurization: On

    –4–
    Unit Handling: Verify
    NaN Handling: Off

    6. Sentiment data. A stock with less sentiment volume (buzz) compare to the average
    sentiment volume of the last quarter (60 trading days) might imply lower risk of
    overraction and is better to invest.

    buzz = ts_backfill(-vec_sum(scl12_alltype_buzzvec),20);
    ts_av_diff(buzz, 60)

    Region: USA
    Universe: Top3000
    Delay: 1
    Neutralization: Sector
    Decay: 4
    Truncation: 0.08
    Pasteurization: On
    Unit Handling: Verify
    NaN Handling: Off

    3.2 Chinese market


    1. Volatility of daily turnover in last 20 days

    rank(-mdl175_volatility*log(volume))
    *(1+group_rank(mdl175_revenuettm, sector))

    Region: CHN
    Universe: Top3000
    Delay: 0
    Neutralization: Sector
    Decay: 3
    Truncation: 0.08
    Pasteurization: On
    Unit Handling: Verify

    –5–
    Figure 2: Results for the price weighted average alpha.

    NaN Handling: Off

    2. Volatility of daily turnover in last 20 days with an event threshold

    # Volatility of daily turnover in last 20 days


    event = volume>adv20;
    alpha = rank(-mdl175_volatility*log(volume))*(1+
    group_rank(mdl175_grossprofit, subindustry));
    trade_when(event, alpha, -1)

    Region: CHN
    Universe: Top3000
    Delay: 0
    Neutralization: Industry
    Decay: 6
    Truncation: 0.05
    Pasteurization: On
    Unit Handling: Verify
    NaN Handling: On

    In Fig. 2 we present the results for this trading signal.

    References

    [1] “Worldquant brain.” https://platform.worldquantbrain.com/.

    –6–

    You might also like