Repaints Stochastics Divergence System For ThinkOrSwim

[Chence27]

An update to the system:

Right now I am only using Stochastics 9,3 and 60,10 plus cumulative delta (cumulative delta divergences only make up less than 1/4 of the signals - TOS users will have to ignore this for now but it doesn't affect the overall accuracy rate anyway).

On stochastics (60,10) I am now using the faster k line (60) for divergences rather than the slower d line (10) because the d line lags way too much and the k line picks up way more divergences. The accuracy of the k line divergences is scary! At least on es.

This system has a very high accuracy on 1 hour, 30 minute, 15 minute and 5 minute charts. Time frames below 5 minutes also have a good accuracy rate but I'm still conducting tests to confirm the exact numbers. I am not using the other confirmation indicators listed in the original post, as I'm now using Sierra Chart.

In the last month, on a 1 hour chart the win/loss ratio is 21 wins/7 losses, on a 30 min chart it's 23/8, on 15 min it's 56/14 and 5 min is 141/56. This testing was very liberal and I took into account any and all divergences, even those that weren't very strong and that I probably would have passed on in real trading. 15 minute charts and higher can expect each win to be at least 10 points on es (1 hour and 30 min charts average 30 points per win, and 15 min chart averages 20 points per win). Anything less than 10 points I counted as a loss. 5 minute charts and lower will have to settle for smaller wins on stoch(9,3) divergences, but stoch(60,10) signals produce much bigger wins.

 

[Tradervic]

An update to the system:

Right now I am only using Stochastics 9,3 and 60,10 plus cumulative delta (cumulative delta divergences only make up less than 1/4 of the signals - TOS users will have to ignore this for now but it doesn't affect the overall accuracy rate anyway).

On stochastics (60,10) I am now using the faster k line (10) for divergences rather than the slower d line (60) because the d line lags way too much and the k line picks up way more divergences. The accuracy of the k line divergences is scary! At least on es.

This system has a very high accuracy on 1 hour, 30 minute, 15 minute and 5 minute charts. Time frames below 5 minutes also have a good accuracy rate but I'm still conducting tests to confirm the exact numbers. I am not using the other confirmation indicators listed in the original post, as I'm now using Sierra Chart.

In the last month, on a 1 hour chart the win/loss ratio is 21 wins/7 losses, on a 30 min chart it's 23/8, on 15 min it's 56/14 and 5 min is 141/56. This testing was very liberal and I took into account any and all divergences, even those that weren't very strong and that I probably would have passed on in real trading. 15 minute charts and higher can expect each win to be at least 10 points on es (1 hour and 30 min charts average 30 points per win, and 15 min chart averages 20 points per win). Anything less than 10 points I counted as a loss. 5 minute charts and lower will have to settle for smaller wins on stoch(9,3) divergences, but stoch(60,10) signals produce much bigger wins.

Just to make sure I understand you, the faster k uses a 10 period, while the slower d uses a 60 period. This seems opposite to what you had originally in your link where the k had a 60 period and the d had a 10 period. With your new change, the stochastics seemed very choppy. Thanks.

 

[Chence27]

Just to make sure I understand you, the faster k uses a 10 period, while the slower d uses a 60 period. This seems opposite to what you had originally in your link where the k had a 60 period and the d had a 10 period. With your new change, the stochastics seemed very choppy. Thanks.

I wrote that wrong. The faster line is the larger number, so k is 60. It definitely looks choppy using the faster line but it's very good at picking up longer term divergences. The 60,10 stoch is almost useless for quick divergences which the 9,3 is good for, although when a quick divergence is registered on the 60,10 k line, it's very accurate and sometimes won't be picked up on the 9,3. I've attached some images to illustrate some 60,10 divergences. It's the bottom green line. Notice how the k line will pick up the same divergences as the slower d line, but much earlier, and the divergences are much more apparent as well.

 

[Chence27]

 

[Chence27]

This isn't applicable to TOS users, but I figure I'd share because it's part of my system. I look at the footprint to check buying and selling pressure to confirm the divergence signals. For example, today at 14:00 there were multiple divergence signals on es. I looked at the footprint chart and the buying pressure was apparent (the cyan blue font indicates a large buying imbalance i.e. much more aggressive buying than selling), as well as a pickup in volume. This let me know the divergence was valid.

 

[Chence27]

I've also recently incorporated a zero-lag MACD with standard settings for divergences, as it seems to pick up divergences that the stoch and cumulative delta indicators don't find. I would imagine that a regular MACD would also be suitable.

 

[Chence27]

This MACD is just what I needed to complete the system. I'm loving it.

 

[Chence27]

Setups have been golden all day. Let's see if this big boy reversal works out.

Cumulative Delta divergence and MACD divergence setting up on the 1 hr

Volume profile at bouncing off the top of value

2, 3 and 5 minute MACD divergences

 

[rad14733]

@Chence27 Based on the number of images you've posted, you're quite fascinated by divergences... Is that your primary trading strategy...??? Just curious... I've never taken the time to look for, nor had any desire to look for, divergences...

 

[Chence27]

@Chence27 Based on the number of images you've posted, you're quite fascinated by divergences... Is that your primary trading strategy...??? Just curious... I've never taken the time to look for, nor had any desire to look for, divergences...

It's not a fascination. It's simply recognition of a leading indicator with a very high accuracy rate. The point here is to share a strategy that works. If you've seen the win rates I've posted and still have no desire to investigate divergences, I'd have to just scratch my head.

 

[Chence27]

I just checked in on es to see how the potential divergence setup I posted early panned out. Sure enough, price dropped nearly 30 points immediately after my post. Works like magic. Combined with the fact that it was bouncing off of a trendline, price was overextended from the EMAs, and the high of the red doji candle held for the next two candles, showing significant resistance, if you didn't already enter the trade around the time of my initial post - red doji - then hopefully the green doji top tail was your cue to get in for the reversal.

 

[Chence27]

All indicators are picking up divergence on the 10 and 15 minute charts. There's a very high probability that price will continue upwards.

One thing to note is that there's usually a significant shift in momentum at the RTH/Globex transition and this can sometimes throw off the signals. It's less of an issue on larger time frames. I believe 15 minutes is high enough to have some confidence in the move. Lower than that might be iffy around this time.

 

[Chence27]

Today was my first day live trading. It was incredible! I used the system detailed in this thread. Had enough of sim and was thoroughly convinced the system is profitable, so I went for it. It was very successful! I traded MES and hit my profit target of 20 points with about a 75% win rate. Win rate would have been better had I not placed a couple impulsive trades, but as I review my trades later I'll be sure to note my mistakes and try to avoid them tomorrow. Today the market was incredibly choppy and there were less setups than usual but even in these conditions the system performed well. Couldn't be happier!

 

[TraderKevin]

Today was my first day live trading. It was incredible! I used the system detailed in this thread. Had enough of sim and was thoroughly convinced the system is profitable, so I went for it. It was very successful! I traded MES and hit my profit target of 20 points with about a 75% win rate. Win rate would have been better had I not placed a couple impulsive trades, but as I review my trades later I'll be sure to note my mistakes and try to avoid them tomorrow. Today the market was incredibly choppy and there were less setups than usual but even in these conditions the system performed well. Couldn't be happier!

Congrats! Sounds like you're doing things the right way and not jumping in too deep too fast... trading monopoly money for a long time to perfect things first. Now lets see if you have the patience to stick with it after the system has a draw down, which I promise you it will

One critique, if you're interested. This is not a system as you've described it, since a system has rules for both entering and exiting trades (whether profitable or losing). Since it's not a complete system (at least as you've described it here), your winning/losing percentages lose most of their meaning. Just my 2 cents.

 

[Chence27]

Congrats! Sounds like you're doing things the right way and not jumping in too deep too fast... trading monopoly money for a long time to perfect things first. Now lets see if you have the patience to stick with it after the system has a draw down, which I promise you it will

One critique, if you're interested. This is not a system as you've described it, since a system has rules for both entering and exiting trades (whether profitable or losing). Since it's not a complete system (at least as you've described it here), your winning/losing percentages lose most of their meaning. Just my 2 cents.

Thanks!

There are rules for entries... a divergence and confirmation on the footprint chart. As for the exits, I'm still working on that. What I know is that winning trades on charts 15 minutes and higher are least a 10 points, and any lower time frames get at least 3 points. I disagree with your comment about the win/loss percentage being meaningless. The amount of the move I profit from does not affect the win/loss percentage so long as I'm taking profit. A small win and a big win are both wins in my book. What I need to to is optimize. For now I'm sticking with minimum profit targets and as I get a better feel for things I'll adjust.

 

[TraderKevin]

Thanks!

There are rules for entries... a divergence and confirmation on the footprint chart. As for the exits, I'm still working on that. What I know is that winning trades on charts 15 minutes and higher are least a 10 points, and any lower time frames get at least 3 points. I disagree with your comment about the win/loss percentage being meaningless. The amount of the move I profit from does not affect the win/loss percentage so long as I'm taking profit. A small win and a big win are both wins in my book. What I need to to is optimize. For now I'm sticking with minimum profit targets and as I get a better feel for things I'll adjust.

Let me clarify a bit. The reason the win/loss percentage is meaningless is because you have not defined your exits if it is a losing trade (so far as you've said at least). This means that you don't know the size of your average losing trade, so you can't compare it to the size of your average winning trade. Therefore, even if you have a 75% win rate, if your average loss is 3 times larger than your average win, you are not profitable. You also have to factor in slippage, commissions, and exchange fees, which will likely eat into your profits significantly. That's why I say your win percentage doesn't matter, as it will only ever matter if you can analyze it in conjunction with your average win and average loss sizes.

As an aside, people focus far too much on having strategies with high win rates. My best strategy only wins about 42% of the time, but the average winner is over twice as large as the average loser, so it is quite profitable.

 

[Chence27]

Let me clarify a bit. The reason the win/loss percentage is meaningless is because you have not defined your exits if it is a losing trade (so far as you've said at least). This means that you don't know the size of your average losing trade, so you can't compare it to the size of your average winning trade. Therefore, even if you have a 75% win rate, if your average loss is 3 times larger than your average win, you are not profitable. You also have to factor in slippage, commissions, and exchange fees, which will likely eat into your profits significantly. That's why I say your win percentage doesn't matter, as it will only ever matter if you can analyze it in conjunction with your average win and average loss sizes.

As an aside, people focus far too much on having strategies with high win rates. My best strategy only wins about 42% of the time, but the average winner is over twice as large as the average loser, so it is quite profitable.

But I have defined my exits, so I do know the size of my average losing trade and can compare it to the size of my average winning trade. I wouldn't say win percentage doesn't matter, because it does matter... but it's not the only thing that matters. Like you said, it's half of the equation.

 

[Tradervic]

I've also recently incorporated a zero-lag MACD with standard settings for divergences, as it seems to pick up divergences that the stoch and cumulative delta indicators don't find. I would imagine that a regular MACD would also be suitable.

Can you shed more light on this zero lag MACD? How did you modify it? Thanks

 

[Chence27]

Can you shed more light on this zero lag MACD? How did you modify it? Thanks

It's a custom indicator for sierra chart. I don't know much about it other than it gives great signals. I compared it to regular macd and it's far superior for giving divergence signals.

 

[TomNevers]

I'm a long time believer in Stochastics Divergence but I've never been able to set up a scan to capture them real time. Can anyone share an idea for how to scan for them? I doesn't sound like it would be that difficult however coding isn't my specialty.

To be clear a divergence occurs when;

1) Bullish Entry: Price makes a new low however the stochastics indicator does not make a new low. The stoch indicator makes a higher low.
2) Bearish Entry: Price makes a new high however the stochastics indicator does not make a new high. The stoch indicator makes a lower high.

Some of the screenshots above have this backward.

Thanks - I'm a first time poster.