January 31 Close Comments

[Mike]

The table below shows the performance of breakouts since the beginning of the rally. The Follow-Through Day was December 20. What we like to see is many stocks showing 20-25% profit or more giving us a chance to meet normal profit targets for a purchase. Only INVN meets this requirement. The rest of the stocks are trying to get there. In a normal CANSLIM rally we should have already seen a sizeable fraction of the stocks in this harvestable position. It appears to me that the indices are looking more like a good rally than the action of stocks with good fundamentals. So this is a tradeable rally so far but one in which it looks like the profit target should be lowered to the 10-15% range instead of the usual 20-25% range. This also would require a change in stop loss on a position to 1/3 of the profit target to keep the risk/reward in the correct relationship.

Even with many earnings estimates having been lowered prior to the current earnins season, the earnings beat rate has been below recent history. I noticed the comment about the Baltic Dry Shipping Index which seems ominous that we could be headed into a worldwide slowdown. I don't know what is levitating this market but just keep watching price and volume.

The Market School Exposure Count is +5 (fully invested), the distribution count on the NASDAQ is +1 (very light).

12546

[EB]

I don't know what is levitating this market but just keep watching price and volume.
12546

It might just be as simple as incremental Fed liquidity, now that Draghi is backdoor bailing out Europe. Having said that, Portugal looks like it will become the next source of headline risk and may eventually lead the markets into the next correction.

12557

[Riskslayer]

Hi Mike,

I was wondering if you could check my work here.

I think Tu (Jan 31) may have been a stalling day on $COMPQ in the Exposure Model based on the rules you provided (perhaps I miss applied them). If there was stalling, I take that its the first stall day in the count and is counted, thus this stall day plus the Th, Jan 26 Distribution Day would bring the Distribution Count to 2?

Here is my analysis (I have restated the rules to make it easier for me to quickly conclude a day is not stalling, thus helping me more quickly analyze suspect say - or, so I hope):

1. Any close >= to +0.40% is NOT stalling; even if there is a severe reversal after a morning gap up.
YES

2. Any close < 0% is NOT stalling (maybe a DD, need to check, but not stalling).
YES

3. Volume is > 95% of previous, else NOT stalling.
YES

4. At least 1 or 2 previous days have a positive close of >= +0.2% (porosity threshold). Thus, if 2 previous days are NEG, then NOT stalling.

Fri, Jan 27, $COMPQX +0.4%

5. Near a intraday price high: (i) intraday high is >= the 2 previous days's intraday highs, or (ii) intraday high is >= the previous days's high & w/n 3% of 25 day intraday high

YES

6. % close in range test: (i) % range > 80%, then NOT stalling, (ii) % range < 50% stalling, else 2 special cases, (a) 50% < % range < 65%, apply special test #1 (narrow range stalling), or (b) 65% < % range < 80%, apply special test #2.

% Range = 48%

Thanks,

Shawn

[Mike]

Shawn,
You have an eagle eye.
Tuesday was a stalling day on the NASDAQ.
I like your restating the rules this way. Indeed we count the first stall day and then we wait for more real distribution before counting another stall day. We would need three real distribution days to count another stall day. I know there was a debate about how to handle the case of price down between 0 and 0.2% which is too small to call distribution, whether it was possible to call any of them a stall day. They said no.

[Riskslayer]

Shawn,
You have an eagle eye.
Tuesday was a stalling day on the NASDAQ.
I like your restating the rules this way. Indeed we count the first stall day and then we wait for more real distribution before counting another stall day. We would need three real distribution days to count another stall day. I know there was a debate about how to handle the case of price down between 0 and 0.2% which is too small to call distribution, whether it was possible to call any of them a stall day. They said no.

Thanks Mike.

So, we have had 3 real Distribution Days on $COMPQ, but 2 have fallen out of the Distribution Count because $COMPQ exceeded the closing values on each of the 2 DD's by +6%. Nonetheless, we still keep the stalling score based on all 3 DD's - interesting.

Intuitively, I would have assumed that we only need to track the DD's that were still being counted in the Distribution Count and not the ones that had fallen out.

Thanks,

Shawn

[Mike]

Actually we count the first stall day regardless and thereafter we wait and don't count another one until the total count will contain more distribution days than stall days (for example 3 distribution and 2 stall days). The uncounted stall days are kept in your back pocket until the time they can be counted or they fall out of the 25-day/ 6% window. One of the consequences of bringing a stall day out of your back pocket is that on that day you will see the distribution count increment by two.

Here is a possible sequence of five days:
*Stall Day, Count = 1
*Distribution Day, Count = 2
*Stall Day, Count = 2 (one in the pocket)
*Distribution Day, Count = 3 (one still in the pocket)
*Distribution Day, Count = 5

Incrementing distribution count by 2 can cause some unusual sell signals such as: S3-AND-S4 or S4-AND-S4 on the same day.

As you have discovered the stall day rules are fairly difficult to use. A spreadsheet helps a lot and can automate the entire distribution counting process.

In counting S4 (full distribution count) sell rules I learned something last week. An S4 Full Count or S4 Full Count+1 or S4 Full Count +2 should be counted only once for each level. So if this were the distribution count sequence:
3,4,5,6,7,6,5,6,7 We would see an S3 for the 5 level, and S4 for the 6 level, another S4 for the 7 level but we would not see another S4 for the second time we arrive at the 6 and 7 levels. If the count falls back to Full Count-2 (4) then we reset the S4 flag and can have them again.