Trading Signal for TSLA - Statistical Odds

OLTO Trading Mentor

1. Signal

⬆ Signal type: Long

🎲 Probability: 59.63%

📝 Name: The candle crossed ema(22) from bottom to top

⚒ Instrument: Tesla (TSLA)

🎚 Timeframe: D1

⏰ Current time (America/New York): 2021-06-15 09:30:31

📌 Current price: 616.5 USD

📅 Optimal maximum holding period: until 2021-08-20 (66 days)

⛔ Optimal Stop Loss: 564.07 USD (3.5 model size)

✅ Optimal Take Profit: 878.65 USD (5.0 stop loss volume)

Left: TSLA chart for the last 3 months. The green dashed horizontal line is the optimal take profit for the signal, the red dashed line is the optimal stop loss for the signal. Right: TSLA chart for the last 23 days, on the basis of which the signal was formed.

2. What does "The last candlestick crossed the EMA" signal mean?

The last candlestick crossed the EMA:

An up signal occurs when the last formed candlestick crosses the exponential moving average from bottom to top.

A down signal occurs when the last formed candlestick crosses the exponential moving average from top to bottom.

Left: "The last candlestick crossed the EMA" pattern generates a Long signal. Right: "The last candlestick crossed the EMA" pattern generates a Short signal

3. Statistical Odds for "The last candlestick crossed the EMA" signal

3.1. Search for the optimal maximum position holding period for TSLA

For the generated signal, we conducted an MFE/MAE test (see definition here) in order to understand the most advantageous position holding time (in days), if this signal is followed. The higher the values of the MFE/MAE curve of 1, the more chances of getting the greatest return in relation to risk for a given position holding time (in days).

The maximum value of the MAE/MAE curve (see the figure below) corresponds to the duration of holding an open position for 66 days.

MAE/MFE curve for the "The last candlestick crossed the EMA" signal for the TSLA instrument. If the ratio MFE/MAE > 1, then the entry point has an advantage and gives a better chance to enter a trade with less risk.

3.2. Probability of a positive signal outcome for TSLA

We found all the same signals in the period from 2010-08-03 to 2021-03-12 (there were 161 of them) and calculated the probability of positive outcomes after 66 day(s) and other characteristics in case of following such signals.

🔄 Total number of signals: 161

🎲 Probability of positive outcomes: 59.63%

📈 Signal average profit: 32.645%

📉 Signal average loss: -12.653%

🍏 Signal average return (expectancy): 14.357%

3.3. What happened to the TSLA price after generating the same signals in the past?

We generated 10 of the same signals in the past and tried to simulate what happened to the price for the next day, the next 2 days, 3 days, 1 week, 2 weeks, 1 month and 3 months (price change as a percentage of the price value at which the signals were generated). From the table below, you can see that 10 signals were generated during the period from 2019-11-29 to 2021-02-02. Green blocks mean that the price has increased over a certain period. Red blocks mean that the price has decreased over a certain period. The percentage of price change is indicated inside the cell.

What happened to the TSLA price after generating the same 10 signals in the past between 2019-11-29 and 2021-02-02?

We also generated 10 trajectories with a duration of 66 days of TSLA price movement after each of the 10 signals was generated (see the picture below). The red line is the average for all such trajectories and describes how the TSLA price has changed on average after generating the same signals in the past.

10 trajectories of the TSLA price movement within 66 days and their average (red line) after generating 10 of the same signals in the past

3.4. Optimal stop loss and take profit for the signal

We tested the TSLA for the period from 2010-08-03 to 2021-03-12 to find the best combination of stop loss and take profit levels for the “The last candlestick crossed the EMA” signal.

We used 2 models for setting a stop loss: by the size of the candlestick pattern, relative to which the signal is generated, and by ATR (Average True Range). The stop loss level is additionally varied by multiplying its size by a coefficient from the interval from 0.5 to 3.5. The take profit level is calculated relative to the size of the stop loss and additionally varies by multiplying the size of the stop loss by a coefficient from the interval from 0.5 to 5.

The criterion for the optimal combination of stop loss and take profit was maximization of the average return on the signal for the period from 2010-08-03 to 2021-03-12.

For the model of setting a stop loss relative to the size of the signal candlestick pattern, we got the following result:

"Stop Loss and Take Profit Test" for the "The last candlestick crossed the EMA" signal for the TSLA instrument using candlestick model size. The size of the optimal distance for placing a protective stop order and take profit from the signal generation point is determined by the maximum value (average return if the signal is executed) inside the table cell.

For the model of setting a stop loss relative to the ATR value at the signal generation point, we got the following result:

"Stop Loss and Take Profit Test" for the "The last candlestick crossed the EMA" signal for the TSLA instrument using ATR. The size of the optimal distance for placing a protective stop order and take profit from the signal generation point is determined by the maximum value (average return if the signal is executed) inside the table cell.

The tests showed that we will get the following optimal parameters

✴ The best stop volume model: model size

🍏 Average signal return: 9.469%

⛔ Optimal stop loss ratio: 3.5

✅ Optimal take profit ratio: 5.0

3.5. Same Stop and Take Profit Test

See "Same Stop and Take Profit Test" definition here: Same Stop and Take Profit Test

For the generated signal, we additionally carried out the "Same Stop and Take Profit Test". We have modeled the percentage of profitable trades when placing a protective stop order and take profit at the same distance from the signal generation point. For testing, we used 2 types of stop order distance from the signal generation point, multiplied by a factor from 0.5 to 3.5 with a step of 0.5: a distance equal to the size of the candlestick pattern and a distance equal to the ATR for the TSLA instrument at the moment of signal generation.

"Same Stop and Take Profit Test" for the "The last candlestick crossed the EMA" signal for the TSLA instrument. The percentage of positive signal outcomes is indicated inside the cells of the table.