This introduction was a bit subjective, and I do not want to stay there, but do my first analysis. I want to compare transaction cost and slippage for the index futures. So this is my question:
How much volatility can I buy with the commission for 1 roundturn and 1 tick slippage (representing the bid/ask)?
To do this, I measured the average range of the 15 min bars over the last two weeks. I excluded the night session, as volatility is too low for trading. And I chose the average range and not the average true range on purpose, because I did not want to include the opening gap, which I cannot trade, as I do not hold overnight positions.
For the US Index Futures the RTH session lenght is 6h45, which is 27 15 min bars. I coded an average range indicator and used a period of 270, which represents the last 2 weeks. For the EUREX traded index futures the session length is 8:30, which would be 34 15 min bars, so I used a period of 340 for my range indicator to establish the average range of a 15 min chart over the last 2 weeks.
Results (15 min range, point value / slippage / commission)
ES range 3.22 points * $ 50,- = $ 161.00 / slippage $ 12.50 / commission $ 4.00
TF range 2.75 points * $ 100,- = $ 275.00 / slippage $ 10.00 / commission $ 3.60
NQ range 6.19 points * $ 20,- = $ 123.80 / slippage $ 5.00 / commission $ 4.00
YM range 26.34 points * $ 5,- = $ 131.70 / slippage $ 5.00 / commission $ 4.00
FESX range 8.68 points * € 10,- = € 86,80 / slippage € 10.00 / commission € 4.00
FDAX range 15.64 points * € 25,- = € 391,- / slippage € 12.50 / commission € 4.00
I now calculate the trading cost as a percentage of the average range result for an average 15 min and an average 5 min trade. For the five minute trade I assume that I obtain the approriate volatility measure by dividing the range through sqrt(3).
Results 15 min trade / 5 min trade
ES 10.25 % / 17.75 %
TF 4.95 % / 8.57 %
NQ 7.27 % / 12.6 %
YM 6.83 % / 11.84 %
FESX 16.13% / 27.94 %
FDAX 4.22 % / 7.31%
This looks quite interesting. The precentage shows, which fraction of a profitable or non-profitable trade that lasts 5 or 15 minute is eaten away by the broker and the bid/ask. Quite a lot. You should not trade FESX or ES on a short term time frame.
To evaluate the impact of this result, let us apply it to a simple strategy with an assumed
- percent profitable 60%
- win-to-loss ratio 1:1
What is the expectancy in terms of the money value of the 15 min range?
ES 0.6 * 89.75 - 0.4 * 110.25 % = 9.75%
TF 0.6* 95.05 - 0.4 * 104.95 % = 15.05%
NQ 0.6* 92.73 - 0.4 * 107.27 % = 12.73%
YM 0.6 * 93.17 - 0.4 * 106.83 % = 13.17%
FESX 0.6 * 83.87 - 0.4 * 116.13% = 3.87%
FDAX 0.6 * 95.78 - 0.4 * 104.22 % = 15.78%
Interesting result, is n't it? For the same strategy FDAX or TF produce an expectancy about 4 times higher than for FESX, and over 50% higher than for ES. As a retail trader, who suffers from the assumed transaction costs and does not need the liquidity of ES for trading size, will you ever touch FESX or ES again?
My present for all scalpers
ES 0.6 * 82.25 - 0.4 * 117.75 % = 2.25%
TF 0.6* 91.43 - 0.4 * 108.57 % = 11.43%
NQ 0.6* 87.40 - 0.4 * 112.60 % = 7.40%
YM 0.6 * 88.16 - 0.4 * 111.84 % = 8.16%
FESX 0.6 * 72.06 - 0.4 * 127.94% = -7.94 %
FDAX 0.6 * 92.69 - 0.4 * 107.31 % = 12.69%
Haha, good news for ES scalpers. The expectancy is still positive, as opposed for FESX. But how positive? Let us look at a small example.
I will take the example of ES. The assumed range for a 5 min bar was 3.22 points divided by sqrt(3), which is 1.86 points or approximately 7 ticks. So now you trade your strategy, which has a profit target of seven ticks and a stop loss of 7 ticks with a win/loss ratio of 60%. The value of the 7 ticks would be $ 87.50, the expected profit will be 2.25% of that value or $ 1.97 per trade.
Who is getting rich here, the trader or the broker?
Conclusions
(1) Results are unfavourably affected by transaction costs.
(2) Never trade FESX.
(3) Don't trade ES, except if you have institutional level commissions and you trade volume.
(4) Scalping is a loser's game with a retail account.
(5) If you plan to hold your position about 5 minutes, trade TF or FDAX.
(6) For trades longer than 15 minutes YM and NQ are also be a also reasonable alternative.
I have made these calculations the first time and rapidly, directly here within the dialogue box, so it is well possible that there are some systematic errors or miscalculations. Also it is late in the evening.
