I know I recently said that BID/ASK data isn't necessary to read the market, using the Wyckoff Method. (I admit though, I watch it closely!) However, it may be a very good way to validate the signals ...see attached. Ideally, any effort to create an indicator that signals "the herd" dumping shares at a low ...should correlate with the real event - for which you need BID/ASK data.
My only thinking here ...is that BID/ASK data would provide the best way to validate the study.
Since this effort obviously has momentum ...and is certainly a just cause ...I thought I'd take retract my disrespectful comment on BID/ASK data!
In the thinkScript Resource Center there's an example of referencing "Data Types" and it uses the Bid and Ask data
plot ask = close(priceType = "ASK");
plot bid = close(priceType = "BID");
However, your example chart looks more like a point & figure (PnF) chart which Charles Dow invented using prices. Wyckoff used PnF charts, but with "Xs" an "Os". I've been trying to translate codes that I've found from other platforms into thinkScript, but there seems very little interest in it. See Point and Figure for TOS. Also see attachments below. For more information for PnF charts visit PnF University.
A free trading platform that has PnF charts for stock is QuoteTracker.
In some forum or trading publication I read that the "Time and Sales" table on TOS is just like data from a ticker tape from Wyckoff's day, but in a different media.
Last edited by StockJock; September 16th, 2010 at 01:21 PM.
What you're seeing, in my example, is Delta Colored Volume. That is the volume at each price, color-graduated by higher or lower Delta. (Delta is the difference between Buying volume and Selling volume.) This isn't available on the TOS platform ...and I run fin-alg's plug-in in NinjaTrader ($59/month).
Check out MarketDelta.com, when you get a chance. I think you'd appreciate their stuff.
Yes, I agree it has some impressive displays with advanced information. Currently, I'm at the trading student intern level and the cost/risk ratio would be very high. When I get to your level of competence, I would consider the subscription fee as a "cost of doing business" and maybe be able to write it off on my income tax as a business expense. I could then afford to use TradeStation or the new Prodigio platform. Until then, I'll keep on my self-directed study of Wyckoff's Trading Method etc., and using TOS paperMoney with free indicators.
It's for automated trading. So, anyone who wants to use it needs a proven strategy that has well defined rules that can be programed. And be willing and able to take the risk of having a computer enter and exit trades for you.
Anyway, we don't want to stay off the topic of discussion too much.
I totally agree. When I first tried out the VPA I turned the alarms on but then quickly turned them off because they were so annoying. I haven't looked at them since. I'll put your suggestion in the hopper for things that should be done.
Wish me luck. Woke up with this idea and I'm getting underway.
- Define the most important elements of the Wyckoff Method
- Define the market conditions under which those conditions would be appropriately signaled
- Identify, within the VPA study script, the elements which indicate criteria met for each important condition
- One-by-one, plot each signal (or signal cluster) onto the chart, i.e. hiding all others
- Gather data to determine the signal's statistical "Sensitivity," "Specificity," and "Positive Predictive Value"
The result of such a study would enable us to make statements such as:
The study's "Test of Supply" indicator is 90% sensitive, 85% specific, and has a predictive value of 82%.
The study's "Buying Climax" indicator is 80% sensitive, 75% specific, and has a predictive value of 65%.
I will limit the study to the ES, 6-Month, 1-Hr chart. All definitions will comply with Wyckoff (Pruden/Williams) Methodology. Script that does not apply to Wyckoff-oriented trading (e.g. "Gravestone Doji" script) will not be included.
For buying, no attempt will be made to signal reversals - i.e. identify bottoms and tops. The probability of correctly identifying reversals is less than 50%, i.e. poor odds. Instead, the study will focus on what Wyckoff traders refer to as "Re-accumulation" and "Re-distribution."
In practical terms, this means the study will attempt to identify key Wyckoff-based buy signals - during price advances, once a buying climax has marked the beginning of "re-accumulation". In other words, this means buying advancing markets, once there is evidence of large speculator accumulation and exhausted supply. (The inverse would be true for declining markets.)
Using Wyckoff method, there is no such thing as a "buy" or a "sell signal." Rather, there are indications of supply and demand dynamics (portrayed by price-volume-spread relationships). The Wyckoff-oriented trader integrates those "tells" into their overall decision-making.
Note: This may flop ...or end up requiring more energy than I can devote to it. So don't get your hopes up.
The following 2 users say Thank You to swimtrader for this post:
You've developed a well defined plan to apply Wyckoff's Method. Best of luck to you. I think every time we try to apply what we learn, the next time we try we can polish it up a little to further improve. Wyckoff didn't seem to like chart patterns or indicators, but for beginners like snowcloud and I they are good to validate our learning process. I think Wyckoff was a genius like Stephen Hawking (theoretical physicist), he did his analysis with purely by mental energy, clear memory and visualization. That's why he is so highly respected and remembered by a large following of traders.
Based upon my hypothesis, I looked first at Supply. If you just had one indicator for supply, given what I said above, you'd want what - in the VPA thinkscript is called "supplyTestInUpTrendBar".
The script describes a relatively low volume, wide spread, down-bar, closing off its low ...in an intermediate to long term uptrend.
That sounds right. And if you're in accumulation, that test should be important.
However, during a 6 month period (over 3500 1-hour price bars), there are only 5 appearances of the signal. 3 out of the 5 were associated with subsequent rallies.
I pulled out the long-term requirement ...no increase in the number of signals. I asked for supply tests in short term up-trends ...got 1 signal.
I saw quite of few supply tests ...some unconfirmed, some confirmed - as you would expect. Visually, however, it was immediately clear that they correlated poorly with periods when the market was attempting to determine if supply had been exhausted and were followed by rallies. They occurred at all levels within trading ranges and advances. So I quickly wanted to focus on just those supply tests connected to an uptrend requirement - as I set out to do, in my earlier post.
Sorry guys. But I'm once again, back where I was. That is ...either you observe the market and interpret Wyckoff principals with your brain ...or you re-write this code ...ground up ...based upon the most critical aspects of Wyckoff-oriented trading strategies. Personally, I'm going with the first. I think it's more reliable ...and it just feels better ...feels smarter.
Please feel free to arrive at your own conclusions ...but I'm going back to work!