Determining a successful backtested strategy?

[keithmj]

I was hoping to get some input from you back testers concerning what you think a successful strategy is.

Here is my scenario, would you consider this a successful strategy?

First of all, I prioritize risk/reward over win rate. Therefore, I was able to create some labels in my strategy that show the R value of the trades. In case you don't know, the R value is just your risk reward ratio. My goal is a strategy that gives me 2R, in other words 1:2 risk/reward. It's pretty similar to expectancy.

Anyway, I had about 25 tickers in a scan and they averaged about 2 trades each using my strategy. So, 50 data points would give some reasonable reliability to the statistics. I typically use a max 10% stop loss on my trades. With that, I found that my P/L was a little over $4,000 with an average R value of 1.21. That's using 100 lot sizes on every trade. I then changed my max stop loss to 15% and found my P/L was over $9,000 with an R value of 1.02. A 50% increase in risk to get 100% increase in P/L seems like a good deal.

The problem is, I'm only getting half my R value goal. Granted, a 15% average gain is nothing to shake a stick at. Also, that 1.02R is more of a baseline of purely mechanical trades. Once I actually put eyes on these trades and make decisions on other factors, such as candle reversals or sector analysis, I may not actually take all the trades, or I may sell before I stop out or hold a strong trade longer that the strategy says for more gains. But on the other hand, I may screw up a good systematic trade once I get my hands on it.

So, I guess what I'm asking is, if this were you, would you consider this a successful strategy? It doesn't meet my 2R goal, but a 1.02R baseline seems like a good starting point. Also, what criteria do you guys and girls look for to determine if your strategy is successful when back testing?

Another note, this is for swing trading. I'm looking to be in a trade for days up to a month or so. This is not for day trading.

If anyone is interested in the code for the labels, let me know and I'll add that to this thread later when I am at my other computer.

Thanks!

 

[HODL-Lay-HE-hoo!]

Hello my good man… I would be interested in said labels. As for your question… I could not answer in good conscious however, what is your strategy and does it repaint?

 

[armybender]

First of all, I prioritize risk/reward over win rate.

This may be the wrong way to think about it. What about probability? Think about the trader's equation.

probability of win * win amount > probabilityof loss * loss amount

 

[JimmyThreeSwans]

Also, what criteria do you guys and girls look for to determine if your strategy is successful when back testing?

I determine a strategy to be successful based on whether or not I can trade it live and have confidence in it. I had several "successfully backtested" strategies that I couldn't trade live, and lost money.

Also, I evaluate based on the "scaling" concept. I don't care how much money it makes per 1R. If you can be consistently profitable (even a little bit) you can scale to large contract sizes and make $$$.

 

[GiantBull]

I focus mostly on my win rate % because I’ve noticed that most of the times when I’m in a trade my risk/reward ratio will change in every single trade. I think it’s quite rare for you to keep it exactly the same in every trade unless it’s automated. I keep a hard stop loss so that never changes only the reward does. Some trades I get out quicker with a smaller reward and other times I hold longer resulting in a larger reward. I usually exit based of support/resistance. Regarding your ratio I would say it depends on the win rate % as well. If it’s below 50% I would say that you will lose long term as that would not be a feasible strategy imo. I personally feel comfortable with a strategy that has a higher win rate % of 65% so the ratio you have would make sense then as more than half the time you will win based on the win rate %. Also, according to Investopedia a good R:R is greater than 1:3

 

[UpTwoBucks]

Since I am a Scalper, I always enter with large share size when volume explodes and the stock looks ripe for buying. This trade happens mostly after a run up then a pullback for the next run up. I never buy the first green candle spike. I'm only in for a few seconds or minutes, so there is no time for a stop loss. My stop loss is a mouse click to get out. I trade premarket and market open, then call it a day. I don't encourage anyone to trade this strategy unless they have many years of practice.

 

[GiantBull]

This may be the wrong way to think about it. What about probability? Think about the trader's equation.

probability of win * win amount > probabilityof loss * loss amount

This is interesting never heard of it before. Is the win amount the number of winning trades or profit you make per winning trade?

 

[armybender]

This is interesting never heard of it before. Is the win amount the number of winning trades or profit you make per winning trade?

No, think about it with this example, assuming you're trading a single ES contract in a long trade in a relatively strong uptrend.

Probability of Win = 60%
Win Amount = bar height + 2 ticks (let's just assume 8 ticks total) = $100

Probability of Loss = 40% (we're in a uptrend - bear is less likely)
Loss Amount = bar height + 2 ticks (let's just assume 8 ticks total) = $100

60% x $100 > 40% x $100
The left side of the equation needs to be higher than the right side of the equation.

If you try to go for more than risk = reward, then you end up with a different equation, but your probabilities might be different. Your stop remains the same, but you now need to estimate a 30% probability minimum and 2X your profit to make it make sense.

30% x $200 > 70% * $100

The probabilities are an estimate, but you can mentally determine these depending on what market you're trading. The idea is that you shouldn't risk more than you can make from a trade - in fact, you should always have at least a slight advantage.