In my last post I said I was going to talk today about win/loss ratio and risk/reward ratio. However, I started thinking about something that ties in with the last post which talks about the emotions of greed and fear in connection with money management. (You can read the last post here: Trading Money Management Greed And Fear). It has to do with 2 bad money management strategies I do not recommend: Martingale System and Kelly Formula.
When you are creating your trading system, you should be thinking logically and methodically plan out how you want to trade. Since we are human, emotions like fear and greed cannot be eliminated completely. But when calmly creating your system, you should not let emotions like fear and greed dictate your decisions.
I know a lot of people reading this blog are interested in automating their trading strategy with the hopes of eliminating emotional trading. After all, an expert advisor does not have emotions and will trade exactly as programmed. Just make sure you are not programming in strategies based on the emotions of fear and greed.
Don’t program fear and greed into your automated trading system.
2 Bad Money Management Strategies I Don’t Recommend
Since money management is arguably the most important part of your trading system, I want to go over strategies to avoid. Besides being risky, these strategies are based on the emotions of fear and greed.
The Martingale System
The Martingale System is derived from a betting system where you double your bet after each loss. The theory is when you eventually win, you win back all your losses plus a small gain. Here is an example…
Let’s say I place a bet for $1 and lose. The next bet would be for $2. If I lose again, the next bet would be for $4. If I win this bet I make $4, which covers the $3 in previous losses and makes me $1. Sounds great, right?
The problem is, an eventual losing streak puts more and more of your money at risk for very little reward. Since your trading account is not infinite, it is only a matter of time before you are risking a large part of your account on one trade. This is not a smart way to trade.
In my opinion, the Martingale System is based on fear. People do not like to lose and have a hard time handling loss emotionally. In theory, using the Martingale System is a way to alleviate the fear of loss by quickly recovering your losses and going into positive again. (Basically, you only have to feel the pain of loss until the next trade where you win back all your losses and make a small profit).
But what happens when you have a string of losses? The fear and pain of loss is magnified with each new loss. So, instead of reducing the fear of loss, the emotions involved in your trading increase with each new trade. In the end, your trading is dominated by emotions, which is not a smart or healthy way to trade.
The Kelly Formula
The Kelly Formula is a statistical, mathematical formula developed by Bell Laboratories in the 1950’s. That’s right, this formula had to do with long distance telephone calls. However, any formula that can be applied to gambling or trading soon will be.
The Kelly formula came from engineering work at Bell Labs.
To use the formula there is some statistical information you need to know about your trading system.
- Percentage of winning trades (W).
- Ratio of average gain of winning trades in relation to average loss of losing trades (R).
Here is the shorthand of the formula: Kelly % = W – [(1 – W) / R]
Now, I’m going to be honest with you. This is already too complex for me to even consider using as my money management strategy. But the real problem is, depending on the numbers you plug into the formula, it is very easy to get percentages of over 20%.
There is no way I would ever recommend risking 20% or more on any given trade. And the only reason I can see to do such a thing would be greed. People using this formula want to figure out the maximum they can risk on a trade so they get maximum returns.
In my opinion, focusing on maximizing your winnings is not the way to survive as a trader. A good rule of thumb is to focus on minimizing your losses instead. I don’t care how mathematically sound the Kelly Formula might be… trading from the standpoint of greed is asking for trouble. Even if using 20% of an ever decreasing account balance never blows out your account completely, this is not smart trading.
Here is the point…
I think money management strategies like Martingale and Kelly are based on emotions. Martingale wants to lessen the fear of loss by quickly allowing you to recoup your losses and go into positive. Kelly wants to maximize your profits specific to past performance so you don’t leave profits on the table. The problem is, real trading is unpredictable and in both cases the wrong circumstances can lead to substantial and quick losses. Plan for the worst and hope for the best.
If you automate your trading and program in money management strategies based on emotions like fear and greed, you entire trading will be corrupted by emotion. The expert advisor will trade as programmed, bad money management included.
If you have any experience with either the Martingale System or Kelly Formula, please leave your comments below.
The Martingale system is based on the most fundamental law of the universe – rhythmic balanced interchange. Mathematically, they call it The Law of Averages. But the bottom line is, every candle that closes bearish increases the likelihood that the next candle will close bullish and vice versa.
The strategy is more suitable for binary options than forex in my opinion though, because you can lose more than you wager when using it in forex.
Hi Terrence,
Thanks for your comment. That’s a common misperception in the markets and in any risk taking endeavor. It’s called the Gambler’s Fallacy.
The Kelly Criterion is actually optimal as can be proven with a simple spread-sheet.
BUT!
The Kelly Bet does not include concepts like “That’s my life savings teetering precariously over there on the hard right edge!”
It has no concept of leverage or margin calls, etc..
So, best bet would be to apply the resulting Kelly percentage to your maximum acceptable loss. Say, if Kelly says to go with 50%, and your max loss is already 2%, then go with 1% in the trade.
Hi Jay,
I hear what you’re saying, but I disagree. You can never know your maximum loss.
When you’re trading with leverage, you live in dread of events like 9/11 or unexpected Fed announcements where the market can gap over your stop. Trading something silly like 10% of your account balance on every trade easily leads to devastating the account.
You’re very right to point out that nobody can emotionally withstand the Kelly induced drawdowns. The formula is designed to sustain losses as close as possible to zero without (theoretically) blowing up. Nobody trading their nest egg, no matter how convinced or deluded about their strategy’s edge, would ever be able to endure those types of losses and to continue.
Hey Shaun!
Thank you for that. And it sounds like we agree more than not. What I meant by “max loss” was what you could walk away from without either emotional or account devastation. The kind of coin that you casually drop off in Las Vegas, just for fun and entertainment. That’s the reasoning that most professional money management systems set a limit of around 1% to 2% of the account balance to be in play on any one trade.
The Kelly formula is kind of basic and over-simplified for leveraged trading. It has no concept of being temporarily wrong and how leverage can stop the trade out before turning back.
If we were to make a graph of the Kelly criteria, basically it forms a cross. On the up-side bigger trade positions make bigger profits. On the down side, bigger trade positions ensure bigger losses. It finds that trade-off point.
But the formula only concerns the outcome. If the trader’s history shows a statistically significant probability of profit, then the formula can show what would have been the best trade position size to take.
So, IMHO, there needs to be some adjustment to the formula to include the temporarily wrong position at leverage. There’s probably no formula for fear, greed, frustration, tears, etc.
So, yes, all the boring trading disclaimers apply.
Well said!
Merry Christmas, Shaun!
Was thinking over the Kelly Optimum problem. Another little detail that few if anyone talks about is that short term results can vary widely from the assumed long term. For example, I’ve just flipped my lucky coin of science 20 times and found that it actually came up tails 13 times and heads came up only 7 times! Including 2 runs of tails occurring 4 times in a row. Therefore, to actually optimize the position size (for real trading where we are limited in the number of times we can enter the market), to ensure that we can stay in the game over these short term deviations we must allow for a good sizable short term variance from the expected long term outcome. Someone must have discussed this, but I haven’t found it mentioned anywhere yet.
Also, what is really meant by a “risk size”? It seems that this is actually referring to the value of the stop loss position. If we can agree that some value of stop loss measured in pips would be appropriate for a certain market condition, then placing a dollar value to that number of pips would be where we look to size the trade entry size.
Thus, if we under-size the trade position, we can move the stop loss inappropriately far while maintaining the same “risk” value.
Over-sizing would cause us to move that stop loss too close and ensure a greater number of stop-outs (invalidating our win/loss estimate anyway)
So, I’m thinking at least be as pessimistic as possible in counting our win/loss ratios. Barely positive doesn’t count. And definitely allow for short term deviation from normal to allow us to stay in the game. In short, full Kelly sizing can only work if we can guarantee that we can continue trading after an unusually long series of losses.
Oh! I love it when I remember stuff!! 🙂
One of the most important reasons for not using full Kelly sized trade positions is that stuff happens, , ,
Like, Could anyone really explain,
Just what did happen on 18 Dec 2013 at just about 21:00 hours?
Looking on the Eur/Usd minute chart, (Actually every currency pair had a similar event at the same time.)
A serious spike up for at least 400 points occurred within one minute, then down from there almost a thousand points within the next minute, back up again a thousand, Down, Up, Down, Etc!
If a short trade were in position at 1.376 and (correctly) waiting for the price to go down, The trade would have to endure going negative for at least 520 points the first half hour and wait, being at a potential loss for over an hour until the price actually went down again and finally being profitable!
Looking over past data, one can find many, many similar spikes just before the price finally moves in the predicted direction!
Seems like the leading edge of these spikes is very often opposite to the final price action.
Thus, I’m thinking that to have a chance against this common action, is to have extremely small trade positions that can have huge stop losses of at least 500 points.
Am I wrong in this ?
Or, consider a worst case scenario, where a trader had placed a sell stop order at 1.37034, Trade would have been activated at that price and would have had to endure going negative almost 1,100 points within the first half hour and then wait it out for almost 2 hours before finally being profitable from then on!
Now how large of a trade position can we place? 😉
The problem with Kelly betting is it requires you bet more during winning streaks and less during losing streaks. Sounds sensible enough. But in reality, you never know when you’re in a winning or losing streak. If you’ve been winning you WERE in winning streak. Anything can happen today. Rarely do we admit we’re IN a losing streak. Why get out of bed?
I don’t see it that way. If your last trade lost, you’re in a losing streak.