Share
410 Posts.
lightbulb Created with Sketch. 697
clock Created with Sketch.
08/10/21
08:27
Share
Originally posted by Semmel:
↑
I am also a software engineer. Today, I had a problem where I worked together with an optics engineer who generated some data for me. I looked at the data and said that it looks odd, too good to be true. But the OE didnt think twice about it and went along his business. It was the first indication that SOMETHING is not right. But he didnt budge at it so I trusted his judgement and went on with the results he gave me. Turns out the data wasnt good. I couldnt do what I was supposed to do because one number was grossly wrong that we both didnt look at initially. So I went to him and said that number is wrong. We found out that the entire optical simulation he used to generate the data was garbage for some small mistake in setting it up. Now we have to regenerate the data. If I see stuff that doesnt fit the entire picture, I get alarm bells ringing. I knew something was wrong when I first saw the data, but I couldnt put my finger on it until I found that other number. Lesson: If some small thing is odd, there is usually a reason. I have this SOOO SOOO often in my job. I get data from all sides in our project and deal with it that the engineers prepare for me. Optical engineers, electical, mechanical .. specialists in detector technology and many other areas where I interface with. And its often me, somewone who has no deep understanding of the specialised fields, takes the stuff of experts apart and have to tell them that what they prepared for me doesnt make sense. The thing is, I look at the data and search for inconsistencies. And these inconsistencies usually lead to deeper problems. For example that one time, the detector we use has 6k by 6k pixels and an area of artificial pixels around the true image space that are used to track the performance of the image readout electronics. These "pre scan" and "over scan" areas are usually a power of 2 pixels, say 32. These pixels are not phyisically there, they only show up in the final image because the image readout process is exercised without actually taking data. This one image had 33 pre-scan pixels at one edge of the detector, but not the others. And you could only see that in the noise pattern, so it was super non-obvious and you woudn't notice that by looking at the image. The detector engineer that gave me that data almost started to believe in divine intervention when I showed him proof of what I found. And coming back to Talga, the share price at moment is an inconsistency for me. If valued correctly and even if taking into account the risks of permits, JV, etc.. Talgas share price should be significantly higher. And that doesnt even include the hype around batteries at the moment. So for me, there are two options left. 1. I missed something and there is more underlying risk than I give it credit. Or 2. that the market is inefficient and there is an investment opportunity. I bet on number 2. But I want to make damn sure that its not number 1.
Expand
"And coming back to Talga, the share price at moment is an inconsistency for me." This is fair. In terms of investing in small caps with low trading volumes, I think SP can often be misleading though. Just looking at TLG these last two weeks, the SP has ping-ponged back and forth between $1.60 and $1.35. That's a huge trading range without any news coming from the company to justify that kind of volatility. (I believe it's a sign that TLG is a coiled spring right now. We just need a catalyst to drive demand.) With small caps, I think your edge is your research. Your research is what gives you the conviction to hold while you wait for the market to catch up to the opportunity - and being ahead of the market is what makes outsized returns possible in the first place. Since the upside is so high, you don't even have to be right about every small cap you invest in - a couple winners will pay for many losers.