Newbie, Jones said start with a one lot. Listen to him, he knows what he's talking about. If you're still around in a year, then you can think about fixed ratios etc. In fact, I would say only trade one e-mini. I know you won't listen, but now I feel better...LOL.
IMO fixed ratio method doesn't yield any significant advantage over fixed fractional. Maybe the only one is that fixed ratio seems to work better for trend following systems (low winning percentages, usually around 40%, but high average win/loss ratio, usually greater than 2) because it decreases faster the # of contracts (shares) when such a system is not in tune with current market conditions, thus reducing probability of large drawdown. This is because systems with low winning percentages are very prone to losing streaks. On the other hand, fixed fractional performs very well for systems with high winning percentages (which I personally prefer) even when the average win/loss ratio is only 1.0 (or even slightly less). High winning percentages plus fixed fractional enables a trader to recapitalize more frequently his account after every winning trade because the probability of several losing trades in a row drops off significantly enabling greater leverage. Of course with fixed fractional drawdowns are usually larger than with fixed ratio but we are traders, not investors and these ups and downs are common in every speculator's life. Good Trading
I just realize one of the major reasons why I lost so much dough trading stocks last year..... position sizing. I have a buddy who subscribes to a chat room where they pound this idea more than they pound the stocks calls themselves. He showed me that you can be -$2.00/share for the day and still make money. It's basic-but enlightening. Here's how: Assume your method you trade is accurate only 35% of the time and the win to loss ratio is 2 to 1. If you risk 1% of you $100,000 account on each trade ($1,000) and you know where your stop is before you enter a trade-you can be profitable. Think about it- if you have these numbers-and make 100 trades-35 winners and 65 losers- You win $5 (35 winners times $2=$70 minus 65 losers times $1). Here is an example below. EXAMPLE: Trade #1 has a stop $3.50 away Trade #2 has a stop $1.50 away Trade #3 has a stop $2.00 away Trade #4 has a stop $.50 away On trade #1-you would buy 285 shares of that stock. ($1000 divided by the $3.50/share stop). If you get stopped out..... you lose $1,000. Trade #2 you would buy 667 shares...if you get stopped out- you lose $1,000 Trade #3 you would buy 500 shares and if you get stopped out you lose $1,000 On Trade #4-you buy 2000 shares and the stock goes up $2/share. You make $4,000 on this trade but lost $3,000 on the other three- so you still come out with $1,000. HERE IS THE KEY: What I was doing before was tradin 500 shares on every trade no matter what. This is how I was losing money On trade #1 my total is -$1750 On trade #2 my total is -$750 On trade #3 my total is -$1000 On trade #4 my total is +$1,000 for a total net of -$2,500...... I thought I had to find a way to be right more often than not and ended up losing my butt...... The first example I showed proves you can be right only 1 out of 4 trades and still make money. (the example is obviously hypothetical and designed to fit this example)...... This is something I just learned and thought a lot of other newbies could relate to and help them with their trading before they lose more dough..... Good Luck
And you adjust position size using what criteria, exactly? If you knew in advance that your trade will be a loosing trade, instead of reducing the size at entry, wouldn't it make sense to either not enter at all, or take the opposite side? I thought I had to find a way to be right more often than not and ended up losing my butt... You don't have to be right more than 50% on your entries providing you have a reasonable exit strategy. voodoo
All this sounds like a fancy variation of the 'double up' method which is about 100 years old. Just keep doubling the bet on an even money proposition and you will win one unit for every winner. It's also called the 'due bet' method in horse racing circles. When I was very yound and naive about such things I thought it was the key to the kingdom. I don't even give it a second thought today. I suggest you do the same.
Actually newbie is not trying to do a progression system or subjecting to gamberl's fallacy (the "due method"). He is trying to describe how he changes position size based on risk rather than % of equity. He defines risk based on stop size. Hence, he is suggesting: shares_to_buy = betsize/risk_based_on_stop rather than: shares_to_buy = betsize/stock_price or: shares_to_buy = constant This works as long as you accurately forcecast risk and can account for performance of different stop sizes (do your tight stops produce the same risk factors as your wide stops).
Looked to me (without straining too hard) like the premise was about transforming a losing performance into a winning one via bet sizing. The only valid use of bet sizing that I have encountered involves the Kelly Criterion, which , simply put, maximizes the use of your bankroll and minimizes the probability of ruin. I don't see how the size of the stop is related to the long term profitability of the trade. No correlation as far as I know, maybe even a negative correlation. The only thing we know is that we'll lose less on individual trades when wrong, not how profitable the trade is over the long haul.
I don't see how the size of the stop is related to the long term profitability of the trade The reason it is related is that different stop sizes have different risk factors. Take a system that exits on lowest-low of last 10 bars. Sometimes the lowest-low of last 10 bars is only X% of the average true range (i.e. measurement of % volatility), whereas other times it will be Y%. What I recommend is that you look at your results and separate all your trades into different buckets of % volatility stoploss. You will see that tight stoplosses may win only 35% of the time with higher profit:loss ratio, where as wider stop losses way win 60% of the time but much smaller profit:loss ratio. The best way to optimize results, via Kelly or optimal f, is to use a different 'f' for each type of trade. In other words, most trading systems that undergo Kelly or optimal f to try and increase overall profitability average the optimal factor over all trades, when in reality if you are using a variable stoploss you are entering several different trades each with it a different risk factor. There was another thread where I went into some detail on applying 'optimal f' to variable position sizing.