PreprintPDF Available

With a pinch of stop loss: a Real Vision trading recipe

Preprints and early-stage research may not have been peer reviewed yet.

Abstract and Figures

This research note is an excerpt of our paper “With a pinch of stop loss: A Real Vision trading recipe”. The paper analyses the sentiment and compares the performance of trade ideas extracted from 292 interviews of professional investors and money managers, which where streamed on the “Netflix of Finance” - Real Vision between January 2018 and March 2020. This short note gives unique insights into the performance of the trade ideas and analyzes the usefulness of using defined targets and risk limits in trade specifications.
Content may be subject to copyright.
Research Note:
With a pinch of stop loss:
A Real Vision trading recipe
Moritz Heidena,
, Dominik Schnellerb, Moritz Seiberta
aMunich Re Investment Partners GmbH, @moritzheiden, @moritzseibert
bUniversity of Augsburg, Chair of Statistics
This research note is an excerpt of our paper “With a pinch of stop loss: A Real Vision trading recipe”. The paper analyses
the sentiment and compares the performance of trade ideas extracted from 292 interviews of professional investors and money
managers, which where streamed on the “Netflix of Finance” - Real Vision between January 2018 and March 2020. This short note
gives unique insights into the performance of the trade ideas and analyzes the usefulness of using defined targets and risk limits in
trade specifications.
Keywords: real vision, trading, text mining, sentiment
1. Introduction
“You have to minimize your losses and try to
preserve capital for those very few instances where
you can make a lot in a very short period of time.
What you can’t aord to do is throw away your5
capital on suboptimal trades.
- Richard Dennis, Market Wizards
Financial analyst recommendations are part of the daily life of
most investment managers. If and under which conditions these
recommendations add value to your investment performance is10
an ongoing debate in academia, see e.g. Jegadeesh et al. (2004)
and Su et al. (2019). Real live trading examples have shown
that actually shorting popular stocks, in this case based on the
recommendations of Jim Cramer presented in the evening TV
show Mad Money on CNBC, can exploit attention-driven buy-15
ing by retail investors, see Niemann (2017).
In this research note, the bar is much lower. Rather than
finding the perfect strategy that makes us rich, we simply want
to know if an investment in a subscription to the “Netflix of
Finance” - RealVision1, can teach viewers (and readers) some-20
thing useful about how to approach trade ideas. RealVision is
an on-demand financial TV channel that conducts interviews
with experts in the field of finance, business and the global
economy. Among them are well know investors, hedge fund
managers, and financial researchers. Content is updated daily25
and accessible via a subscription based oering. In addition
to Real Vision’s own website, content is available via various
Corresponding author
Email address: (Moritz Heiden)
1Real Vision Website
other platforms, such as Yahoo Finance, Refinitiv, Trading View,
and Interactive Brokers.
As of July 2020, according to Real Vision, the channel has30
tens of thousands of paying subscribers and 400kfree subscribers
on Youtube and via its podcast. The average viewer is 38 years
old and 43% of viewers work in the financial sector. 60% con-
sider themselves active traders. 40% of registered users manage
portfolios of over $500k and earn more than $200k per year.35
80% of users have at least a Bachelor’s degree. Hence, similar
to Real Vision’s guest list, the average user is hardly “average”,
but part of a well-earning, well-educated class of financially ac-
tive and specialised professionals.
The variety of topics ranges from daily briefings to regu-40
larly recurring interviews and topics, such as investment ideas
or macroeconomic analyses. We focus on the format called
“Trade Ideas” which was updated frequently between January
4th, 2018 and March 12th, 2020. In 2020, Real Vision changed
some of the formats and as of the date of writing this note,45
“Trade Ideas” has been featured less frequently on the channel
as broader economic topics started to dominate the discussion.
Real Vision states that these interviews should be interpreted as
ideas, not investment advice. In our study, we do not want to
impose our discretionary views; rather, we seek to judge ideas50
that could have been fully implemented based on the informa-
tion provided by the interviewee. Since many of the ideas entail
a very detailed specification of the trade we are able to analyze
this “systematic” version of the trade. We are interested in the
ideas that have at least a defined instrument and trade direction55
(long/short). If available, we include additional information as
pointed out in section 2.1. Note that we do believe that most
likely a better implementation of the trade could be achieved if
additional information of the trade idea were processed by an
informed human trader. For our analysis however, we delib-60
erately keep the discretionary choices to a minimum to avoid
hindsight bias.
2. Data & Methods
2.1. Real Vision Data
We download transcripts of the interviews from January 4th,65
2018 to March 12th, 2020. This period includes 292 interviews,
from which we extract 311 trade ideas. Some interviews con-
tain more than one idea while others contain only broad rec-
ommendations or tactical tilts. The interviews are mostly struc-
tured around the topic of the trade idea and in many cases, this70
trade idea is summarized in the last paragraph of the transcript,
including the instrument, possible entry and exit points, stop-
losses or performance targets at which the trade should be ex-
ited. We only extract an idea if an instrument and a trade di-
rection can be clearly identified, firstly by our text analysis al-75
gorithm and secondly, through a cross-check by a human2. We
automatically separate key information such as the title, partic-
ipants, interview date and publication date. The methodology
also cleans the text based on common natural language process-
ing techniques, and prepares it for a sentiment analysis which80
is part of ongoing research. The algorithm also matches the in-
terview against a dictionary of asset classes to obtain keywords
and topic classification.
Since not every trade specification is always unambiguous,
we stick to simple rules for specifying the trades:85
if no entry price is explicitly mentioned, we assume the
trade should be entered directly after the publication date
of the interview.
if a range of entries, exits or stops is provided, we take
the mid price of these intervals.90
if multiple time frames for the trade are mentioned, we
take the shorter one.
if no instrument or clear direction of the trade is given,
we do not include the trade idea in our analysis.
These rules do not apply if the interviewee clearly states within95
the interview that the ideas should be treated as separate trades.
We currently exclude:
trades that use trailing entries or stops.
trades with an end date after the current date of evalua-
tion, July 1st, 2020.100
trades using options, futures and government bonds.
We mark the trade as having a valid “exit rule”, if it has at
least one of the following
a target price or a percentage target return.
a stop-loss.105
2Yes, we did read all the transcripts.
a defined end-date, which we call time stop.
If a trade is not stopped out by one of the exit rules, it is
stopped out by the current date of evaluation which we mark as
max. date.
Overall, we are left with 282 valid trade ideas. This is fur-110
ther reduced to 269 trades, as 13 trades do not reach their entry
level. Of these ideas, 121 are for Equities (single stocks), 60 for
ETFs, 53 for Currencies and 35 for Indices. For simplicity, we
assume that the mentioned indices are tradable.
Table 1 shows the 10 most common topics of the trade ideas,115
as well as the most common instruments that were mentioned.
Note that since we classify the trade according to the general
topic of the interview, there can be more uses of the instrument
than counts of the topic. For example, for an interviewee who
has an explicit view on the USD and decides to implement this120
view against the EUR, we classify the topic as “USD”. The used
instrument will nevertheless be the EUR/USD.
Count Count
Topic Instrument
Technology 40 SPX Index 19
Large Cap Index 39 XAU/USD 15
Precious metals 29 EUR/USD 9
Financials 10 EEM ETF 7
Agriculture 10 SPY ETF 4
Entertainment 9 KRE ETF 4
Healthcare 8 Disney stock 4
Euro 8 GLD ETF 4
Real Estate 8 Apple stock 3
Energy 8 USD/JPY 3
Table 1: 10 most common topics and instruments.
2.2. Market Data
Market data is obtained via Bloomberg for the period begin-
ning with the publication of the respective idea until the current125
date of evaluation, July 1st, 2020. We use closing prices and
currently exclude dividends in our calculations.3For each in-
strument not denominated in USD, we check which currency is
referenced in the interview and convert prices and trade specifi-
cations into USD using 5pm New York time Bloomberg fixings.130
In addition to our topic classification derived from the pure text,
we also obtain security class information from Bloomberg for
each instrument.
3. Trading Idea Evaluation
For each trade, we calculate the log-return, volatility, and135
average annualized Sharpe ratio (rf=0). For a first analysis,
we examine Sharpe ratios classified by the existence of entry
and exit rules, see Table 2. Of the 269 trades that are entered,
we cannot compute valid volatility estimates for 15 trades since
their lifespan is smaller than two days.140
3This issue will be addressed in the final paper
avg. ann. Sharpe ratio n. obs
Entry Exit
7 7 -0.1272 34
7 3 0.0399 139
3 7 0.3066 7
3 3 0.3316 74
Table 2: Average annualized Sharpe ratio grouped by entry/exit combination.
Entry means a a precise trade entry is specified. Exit means a precise exit
condition is specified.
We observe that while the majority of trades has no defined
entry rule, an exit rule is present for 212 of the trades. Based
on the Sharpe ratio, trades without an entry and exit perform
worst (0.1272), while trades with entry and exit perform best
(0.3316). The number of trades that have an entry, but no exit145
is small, but their Sharpe ratio is surprisingly high. Without
an exit rule, these 7 trades are entered on publication date and
stopped out at the evaluation date beginning of July 2020. It’s
worth to have a separate look at these trades, as three of them
are coming from the same interview. These three trades alone150
have an average return of 42% and are all technology-related in
the single stock space. The second best trade in this group is a
long Gold position with a return of 28%.
Figure 1 shows the relative frequency of the three most fre-
quent topics according to each of the four groups. Trades from155
the technology sector are relatively less prevalent in the other
groups, where on average they make up 15% of the trades vs.
57% in the defined entry but undefined exit group. Hence, the
results in this group might be skewed due to a sector focus.
Figure 1: Relative frequency of the three most frequent topics by entry/exit
Instead of the average of the individual trade Sharpe ratios,160
a clearer picture of the trade characteristics can be provided
based on the distribution of trade returns in Figure 2 and the de-
scriptive statistics in Table 3. Note that this allows us to include
the additional 15 trades with very short holding periods, where
no volatility estimate could be calculated. Hence, the number165
of observations does change slightly. We can see the outliers of
the group with defined entry but no exit. Both groups of trades
with entries exhibit positive skewness, while the groups without
entry rule show negative skewness, pointing towards the exis-
tence of outliers. Trades with defined entry and defined exit170
have the lowest volatility and highest kurtosis among the four
groups, which means these trades have a higher probability of
an upside gain.
Figure 2: Distribution of trade returns grouped by entry/exit combination.
Entry 3377
Exit 3737
n. obs 79 7 149 34
min -0.35 -0.198 -1.74 -0.806
max 1.008 0.862 1.021 0.529
µ0.018 0.211 -0.025 -0.073
σ0.032 0.156 0.081 0.066
skewnews 2.225 0.566 -1.444 -0.264
kurtosis 10.49 -0.988 9.401 1.105
Table 3: Statistics of distribution of trade returns grouped by entry/exit combi-
Table 4 highlights the average trade return based existence
of entry and exit, split up into the three dierent combinations175
of price based exits that the trade can have: a price target, a
price target combined with a stop-loss rule and a stop-loss rule
We observe that having only a stop-loss rule yields a nega-
tive average return. However, these groups in total have only 3180
observations, which makes the significance of the results ques-
Having a price target and a stop-loss on the other hand yields
a positive average return, independent of the existence of an en-
try. The average price target is 1.7% higher than the average185
stop-loss in the groups where both exit rules exist. We find that
4We converted return targets into price targets for this classification.
avg. trade return n. obs
Entry Exit Exit rule
7 7 None -0.073 34
7 3 Price target -0.056 89
7 3 Price target & stop 0.024 58
7 3 Stop -0.081 2
3 7 None 0.210 7
3 3 Price target 0.095 15
3 3 Price target & stop 0.002 63
3 3 Stop -0.177 1
Table 4: Average trade return conditional on entry and specific exit rule.
55% of the trades with a stop-loss and price target hit the stop-
loss. Overall, the probability of hitting a stop-loss is nearly
three times higher than the probability of hitting the price tar-
get. Due to their asymmetric specification, these trades do ef-190
fectively cut losses while maintaining upside potential.
For trades that only have a price target, the specification of
an entry seems to be a decisive factor. Trades with defined en-
try and a price target yield an average return of 9.5% per trade,
while trades without a defined entry and a price target results in195
an average return of 5.6%, see Table 4. Intuitively, this makes
sense: even though the lag between filming and publishing the
interviews is 3 4 days, entry prices and conditions may have
changed in the meantime. While 75% of trades with an entry hit
their price target, only 29% of trades without an entry hit their200
price target. More than 50% of the trades without an entry are
stopped out by reaching our evaluation date. 3 of these trades
have a time limit beyond the evaluation date and are forced to
stop before reaching a designated target date. Hence, for the
rest of the trades within the group, performance is heavily de-205
pendent on the evaluation date. This is less prevalent for trades
with an entry and a price target, where we find that only 7% of
trades are stopped out by the evaluation date.
Trades with undefined entry and undefined exit show an av-
erage return of 7.3% per trade for a total of 34 observations210
and are stopped by the evaluation date in 63% of the cases.
4. Conclusion
In this research note, we analyzed the performance of 269
trade ideas aired on the on-demand financial TV channel Real
Vision between January 2018 and March 2020. Our first re-215
sults indicate, that the return distribution of trade ideas with
a systematic set of trading rules such as entry, exit and stop-
loss exhibits higher positive skew, less volatility and fewer out-
liers compared to trade ideas that are less specific. Especially
for trades with a price target, the specification of an entry rule220
seems to be a decisive factor for trade performance. Overall,
our analysis shows that even for discretionary ideas, system-
atic rules play an important role. Besides making a trade idea
testable, they reduce downside risk while allowing for upside
Jegadeesh N, Kim J, Krische SD, Lee CMC. Analyzing the analysts: When do
recommendations add value? The Journal of Finance 2004;59(3):1083–124.
Niemann M. Essays on Prediction and Behavior of Financial Markets. Ph.D.
thesis; University of Augsburg; 2017.230
Su C, Zhang H, Bangassa K, Joseph NL. On the investment value
of sell-side analyst recommendation revisions in the UK. Review of
Quantitative Finance and Accounting 2019;53(1):257–93. doi:10.1007/
s11156-018- 0749-y.
ResearchGate has not been able to resolve any citations for this publication.
Full-text available
This study conducts a comprehensive investigation into the investment value of sell-side analyst recommendation revisions in the UK, using a unique dataset from 1995 to 2013. Our rolling window analysis shows that, on average, upgrades fail to generate any significantly positive abnormal returns in any period of time, even before transaction costs. In addition, although downgrades could generate significantly negative abnormal gross returns over some periods of time, these observed significant returns disappear after accounting for transaction costs. Overall, our bootstrapping simulations confirm sell-side analysts’ lack of skill in making valuable up/downward revisions to cover the size of transaction costs, irrespective of whether these revisions are made by high-ranking brokerage houses or not. However, an industry-based analysis shows that, within two high-tech industry sectors, i.e., Health Care and Technology sectors, sell-side analysts possess certain skill in making valuable downgrades over some periods of time and, in particular, such skill is sufficient to offset transaction costs.
This cumulative dissertation develops and applies methods to predict and empirically study financial market behavior. It presents three papers examining different research questions on the economic and statistical laws governing financial markets. The first study, Improving Performance of Corporate Rating Prediction Models by Reducing Financial Ratio Heterogeneity, develops a methodology to construct better performing models to predict credit default rates of large corporations across different industries. It was motivated by the fact that our consulting team had difficulties to construct rating models for large corporates, due to limited available data on defaults and heterogeneity in financial ratios across industry groups. Published work did not provide much methodological help. This motivated developing our own methodology to account for industry heterogeneity within the rating model, and thereby achieving a notable improvement in prediction accuracy. The second paper, Exploiting Attention-driven Mispricing: Evidence from Actual-Dollar Trading, develops a systematic trading strategy for U.S. stocks and successfully trades it in a true out-of-sample test with real money. These results not only motivated investors to provide the seed funding to start a quantitative asset management firm, it also posed the question of how these profits could be possible and persistent for a longer period. Given that the widely accepted efficient market hypothesis (Fama, 1970) implies that financial markets eliminate such profit opportunities quickly, this conflicting observation deserved further investigation. Third and finally, the essay High Frequency Trading Intensifies Intraday Extreme Events in Stock Returns investigates whether high frequency trading (HFT) activity exacerbates large intraday price moves in the stock market. The idea of investigating the link between HFT and intraday extreme events was motivated by my intraday market observations from countless hours of automated trading surveillance. Thereby, sudden bursts of activity and volatility – often without any news – were a surprisingly regular phenomenon. At the same time, there is a dichotomy in the literature. On the one hand, several published empirical studies indicate that HFT activity dampens volatility and improves market quality. On the other hand, theoretical models and institutional traders formulate multiple plausible mechanisms by which HFT could cause extreme events in short-term stock returns.
We show that analysts from sell-side firms generally recommend "glamour" (i.e., positive momentum, high growth, high volume, and relatively expensive) stocks. Naïve adherence to these recommendations can be costly, because the "level" of the consensus recommendation adds value only among stocks with favorable quantitative characteristics (i.e., value stocks and positive momentum stocks). In fact, among stocks with unfavorable quantitative characteristics, higher consensus recommendations are associated with worse subsequent returns. In contrast, we find that the quarterly "change" in consensus recommendations is a robust return predictor that appears to contain information orthogonal to a large range of other predictive variables. Copyright 2004 by The American Finance Association.
On the investment value
  • C Su
  • H Zhang
  • K Bangassa
  • N L Joseph
Su C, Zhang H, Bangassa K, Joseph NL. On the investment value