Interview with Covestor Model Manager Frank Hogelucht (SPXU, SDS)

Frank Hogelucht, a relatively new model manager to Covestor (Frank’s Trading the Odds model has an inception date of August 13th, 2010), has an interesting approach to model management:  he combines statistics and historical market data to create a ‘card-counting’ method of picking equities and funds for his model. We spoke with him this week to get a better understanding of his strategy.

Covestor Live: Your model looks for patterns in the market that afford a “tradable edge” historically. Can you give us an example of one of these patterns and the expectations you have based on historical data?

Frank Hogelucht: There’re ‘seasonalities’ like the FOMC announcement session (the day the Federal Open Market Committee announcement for the respective month policy meeting is set to be released).

Historically a strong edge on the long side of the market is provided between the close of the session immediately preceding the FOMC announcement session, until at least the start of the last hour (3:00 pm) of the FOMC announcement session. Since 1990, and out of 165 occurrences (FOMC announcement sessions), the S&P 500 showed a positive performance during the respective time frame on 109 occurrences (or 66.06% of the time), and a median winning trade of 0.61% compared to a median losing trade of -0.37%). A statistically significant positive edge.

On the other side there’re patterns based on price and breadth (like NYSE TICK, the ratio of NYSE advancing vs. declining issues, among others).

For example the ‘bullish hammer candlestick’ (after a bearish sell-off a rally at the end of the session brings price back up into the upper range of the daily trading range) is regularly regarded as a buying opportunity (a strong likelihood of a bullish reversal). But history suggests otherwise: When the S&P 500 posted an intraday low at least -1.5% below the previous session’s close during the session, but closed in the top quartile of its daily trading range, the S&P 5000 was trading lower at the start of the last hour on the then following session on 54 out of 89 occurrences since 1990 (or 60.67% of the time), and closed lower (again) on 51 occurrences (or 57.30% of the time).

So it is regularly better to double-check in order to dissect the rumors from the facts.

CL: One of the positions you’ve recently shifted in and out of in your model is ProShares UltraPro Short S&P 500 ProShares (SPXU). What patterns, data or statistics do you look at to determine when to buy or sell this fund?

FH: SPXU is the exchange symbol for ProShares UltraPro Short S&P500, corresponding to triple (300%) the inverse of the daily performance of the S&P 500. I regularly utilize leveraged ETFs (Exchange Traded Funds) for intraday and short-term trading purposes (at the moment Covestor does not allow for trading futures) when a strong edge on any side of the market is provided (and taking leverage is justified).

CL: On October 18th, you bought and then sold ProShares UltraShort S&P 500 ETF (SDS). Can you share what factored into these decisions?

FH: On Friday, October 15, NYSE TICK data (market breadth) was heavily lopsided in favor of significantly lower prices, but the S&P 500 bucked the trend and closed higher on the day (a rare occurrence). But historical probabilities and odds suggested a downside edge for the then following session (October 18th), which I tried to take advantage of by buying SDS (ProShares UltraShort S&P500). But the supposed move lower never materialized, so I closed the position shortly thereafter in order to keep losses small.

CL: You equate trading in your model to counting cards while playing poker. How do you “count” the economic and financial indicators that can affect stock and ETF prices since they, unlike cards in a deck, don’t come in fixed or consistently predictable quantities?

FH: Some things will never change: greed and fear, together with herd instinct, are the three main emotional motivators of stock markets. Day by day there’re interesting pattern in the market’s (based on seasonalities, price, breadth and sentiment) which might provide a tradable edge for (or over the course of) the next trading day(s), based on a historical and statistical analysis (quantifiable research) how those pattern played out in the past.

Historical probabilities (e.g. winning percentage) and odds (e.g. profit factor, the median winning trade compared to the median losing trade) – like the pot odds in poker – will then determine what (the security or index), when (timing), where (e.g. a leveraged ETF), why (the pattern) and how much (position sizing, leverage) to bet on the trade.