lano1106 UMSv2

[lano1106]

cycle 3. Stage 1 day 6:

My UMS journal starts to reads like a DMSI journal but sexual fulfillment appears to be important to me currently...

I got 2 sex proposals falling on my lap today out of the blue... I invited to my place one of the prospect and this resulted into great and amazing sex... She did stay for about 3 hours...

It is too early to say if this encounter will result into a relationship of any sort...

No idea where all this attention is coming from... It is a bit unusual to me to receive all this attention... I must have changed something in my own sexual marketing without being aware of it...

It is funny that the simple intention of running DMSI without actually run it is bringing me all the action..

that being said... I had a hard time to concentrate on moving forward the exchange migration project... I did move it a little bit forward but it is not fast enough for my taste... I suspect that there is some unknowns in the design that instead of confronting directly the difficulty, I perform some sort of procrastination instead... but I feel that unconsciously, I am organizing my ideas in my mental space... so I have hope that soon, my productivity will unblock once the mental design picture gets clear in my mind...

For one thing, since it is my first framework reuse, I am realizing that its current design lack of modularity. There is a lot of dependencies between the different components... I am not too sure if I can improve that aspect because there are real dependencies between the modules... It is not due to sloppy design... but it annoys the heck out of me because in order to make my next test, I needed to port a module which depends on another one and when the port of the second module is completed, I discover that this one depends on a thirds one and so on...

Without all those dependencies, I could rapidly write one module and try it out... now I need to port the whole thing before making any test at all. The best analogy is getting your finger stuck into some gears and the gears end up ripping your whole arm...

time spent on creating a new exchange layer is not wasted... Everything that I write will end up being used... It is just that more needs to be written before being in position to test... which is not very agile methodology friendly...

Writing this entry did help me figuring out a way to reduce the dependency issue... by creating some class hierarchy...

I could have a BaseExchangeContext with minimal functionality and a FullExchangeContext offering the full API... I am not sure that there would be a use to start passing those 2 interfaces around but just for the sake of testing faster partial implementation, it might be worth the effort... I'll think about it....

[lano1106]

cycle 3. Stage 1 day 7:

I have progressed a little bit again today. It seems like it could be faster but it could also go slower too...

I have the feeling like what I am doing is extremely difficult.
In reality, it is tedious and boring but difficult, it is not...

I try to remember times where I did similar tasks amazingly fast... I try condition myself seeing what I am currently doing easy... maybe if I convince myself of this, it will actually help me doing it faster...

At least, I hope so... I think this could work...

Something that I have realized is that Exchange APIs not of equal quality... The one that I am currently working on contains many small inconsistencies...

ie:
- the same field have different names in different msgs.
- The same field appearing several times in the same msg in different subsections.
- many fields that the documentation says that they will always have a specific value. Therefore, they are useless!
- Fields that are actually helpful to have that are present in comparable msgs in the exchange #1 are missing. This makes the messages almost useless as with the missing info, you will need make other calls to supplement the received info and get the missing one...
- the same fields in different messages or sometimes in the same message use different types. Sometime, it is passed as a string, some other times as an integer... There seems to be no logic at all in the choice. One has to wonder if they did make random choices for the field type.

Seeing all those issues makes my mind try to find workarounds and this is taking time... At some point today, I have made a conscious decision to switch off the critical thinking mindset and turn on the stupid switch... Just implement code to handle the various messages without thinking about the various glitches present in the interface...

[lano1106]

cycle 3. Stage 1 day 10:

Progress is painful... I am not sure why... For one, I have a small cold that doesn't want to go away. Running MIR could help but I resist pausing UMS for doing that...

or it could be procrastination or resistance to something... anyway I am going to break through it with my will force...

A milestone is about to be reached. Hopefully, this event is going to be beneficial to help me getting faster to get things done...

[lano1106]

cycle 3. Stage 1 day 11:

milestone reached.

I bought some books on Amazon the other day. one of them appears to be very interesting as it is precisely on the topic that I have been discovering on my own for the last 2 years. When you work in a new field on your own, the concepts that you manipulate, you do not know the name or the label used for those concepts in the industry.

I have found the field that I have been working on without knowing what it was. It is called Market Microstructure...

It is amazing, I am currently reading the first chapter and it is really cool and really interesting to get a formal academic treatment of what I have been doing for the last few years. My experience is helping me to have clear understanding of what the book is talking about and I think that I am going to discover several ideas directly applicable on my system... For one, I have a great interest in exploring market making opportunities. I feel like my software is a perfect fit for doing market making...

Also, today, out of nowhere, the system got triggered like crazy and a new feature that was yet to be tested has been used successfully. My system used to be edge triggered. That is, if I detect a trade opportunity and my system execute it. Once executed, it was waiting for the next detected opportunity. Now, it checks the available volume in the opportunity and calculate how many times I can execute it with my trading budget. If it can loop through the same trade 5 times, it will attempt to do that.

and it works like a charm. That was the feature that I was hoping to leverage when my system was trading like crazy back in June.

but bottomline, out of the blue... I got hundreds of trades executed with very juicy profit... I suspect that next week-end will experience some big moves... Previous times that my system did perform like crazy, it was a precursor to big moves in crypto market... This is what I am predicting... Lets see if my market prediction skill has become sharp...

Update:
I did figure out what stimulated my system. My system did benefit from this:
https://www.coindesk.com/business/2022/0...-shoutout/

[lano1106]

cycle 3. Stage 1 day 12:

I might journal further later today... The luck effect where I witness a perfect storm showing me flaws in my system did occur again. I woke up with a crash this morning

[lano1106]

cycle 3. Stage 1 day 14:

The more the support task of the second exchange progress, the less it looks promising... It advertise having a bunch of pairs but in reality when you start using their API to get more info about those pairs, you discover that few of them are in a state that basically is equivalent of being unavailable.

Many of the pairs that I was planning to use for arbitrage are in this state. So this makes the exchange potential much smaller than expected... I'll possibly reach completion because the experience itself is valuable and allows me to improve the framework to be more flexible and on top of that, this will open arbitrage possibilities between 2 exchanges. This is something that I have never done before and I suspect that this might be easier to pull than what I currently do...

That being said I am very tempted to put the migration effort on hold because reading my new market microstructure book gave me a new strategy idea that could be faster to deploy and faster to generate profit than supporting a second exchange...

Being fast to generate income has become an important criteria to decide on where I spend my time and energy.

I am now on chapter 3 of my new finance book. First 2 chapters were fun to read but now, it has become a math book with a lot of statistics and probability notions. Last time that I have done such type of reading goes back to university. Because of that the book has become cryptic... I did get my old university books on the topic to refresh my mind on those various notions...

[lano1106]

cycle 3. Stage 1 day 16:

I had to put on hold trading project to manage all the other boring chores...

I am getting done over all those chores and I'll be hopefully ready to continue working on the project soon.

Out of the blue, my system traded about 200 profitable trades...
This makes me wonder if the universe is trying to communicate me to not neglect the other chores...

[lano1106]

Cycle 3 Stage 1 day 25:

My project has been more or less stalling since the last entry.

I now know why I ignore almost everything to focus on moving forward my project. Reading and replying to emails is draining so much time that could be used to advance the project.

I am taking a day or 2 off from emails to focus on the project...

It is funny. As soon as I took this decsion a small glitch has blown off in my face not once but twice... It seems like UMS luck amplification has restarted its magic. I have also found a good way to optimize some code that is in the hot path. I am eager to see the result once this optimization is tried out. It could make a small difference in my system reaction time.

I have put the exchange migration effort as I got an idea that I got from my new book. I have started to write code to put the idea in place. In fact, this is by reworking some code for this exact purpose that I did stumble into the optimization opportunity.

I hope to start having some visibility on the new idea potential maybe tomorrow and if it looks promising, this might be what I need to kick myself and speed up the implementation of it...

[lano1106]

Cycle 3 Stage 1 day 27:

Taking care of last run discovered problems has taken more time than expected.
However, as a result, the codebase is now more robust and is also generating better traces so in my book this was well invested time.

Now, I can finally return to the possible next money generating feature...

[lano1106]

Cycle 3 Stage 1 day 31:

last day of stage 1. The third iteration through stage 1 has been very quiet... By experience, work performed by stage 1 is under the surface... I have always experienced my major breakthroughs during stage 2 and 3... So with this in mind, I'll keep my enthusiasm.

I felt tired this week... Maybe it is because I have a new schedule. I wake up about 1 hour earlier to help the kids prepare for going to the school...

With this difficulty mixed with perhaps some procrastination as I am feeling drifting a little bit, I have kept moving forward my project...

Some results started to get in and it is very promising... My system is flooded with an abundance of new trading opportunities. In fact, too much opportunities. I have started to add even more qualifications to reduce the number so that my system focus only on the highest quality opportunities...

Part of the process, I had reenable a part of my system that was basically not used long time ago and it was not compiling anymore. I had to fix compile errors and I am currently debugging the feature that simply does not work anymore. It makes me review code that I have not worked with for a very long time... Bottomline, all these issues should be resolved sometime tonight or no later than tomorrow morning...

On a different topic, I have started to review my old statistics and probability math book which is basically a prerequisite to be able to read and understand my latest finance books. This effort is going well... Long forgetting notions are starting to resurface and I am amazed by how more motivated you can be to study when you do it with a concrete purpose...

I feel like stage 2 is lining up to be a good month... It is starting tomorrow!

[lano1106]

Cycle 3 Stage 2 day 1:

It took most of the day but I have finally found my problem.

It was hard to find and at the same time a silly problem...

I went from having too much opportunities to having a too narrow filter... I still need to perform some tweaking but this new thing is going to work...

[lano1106]

Cycle 3, stage 2, day 4:

I have completed the task that detects opportunities for my new strategy. The result is quite satisfying.

I got distraction from 2 interpersonal issues... Those have drained my focus...

I am starting to get over it and I have started to be able to focus on the work. What remains to be done is to code that will actually execute the strategy. There is not much to do since it will heavily reuse the code already existing...

I am starting to find ways where I can simplify implementation to make the initial release fast. The more sophisticated more polished implementation is not really needed at first consideiring the amount of activity currently experienced by my system...

As usual, as I am starting to have a good development cruising speed, something is crashing my live system. Hopefully, something trivial happened. The exchange appears to have behaved in a new way that did trigger a crash on my side... Lets call that "luck" again...

[lano1106]

Cycle 3, stage 2, day 6:

Yesterday, I wrote down optional features and this did help me to move forward completion by writing down small aspects design details.

I have been stuck however this morning working on my latest features... Not sure exactly why... I have been drifting, searching for distractions...

and then suddenly... I am not even sure what did cause that but crypto did experience a mini crash... Volatility went through the roof and my system started to trade like crazy. It made something like 3000 trades in 5 minutes. It did trade a volume of over $100K in 5 minutes... It made about $400 of profits again in just 5 minutes... That is for that type of thrill that I have been working my ass off for now almost 3 years...

Overall things went very well. I think that I have only spotted 2 small glitches that I did fix and will release later when I upgrade my server...

[RTBoss]

Cycle 3, stage 2, day 6:

Yesterday, I wrote down optional features and this did help me to move forward completion by writing down small aspects design details.

I have been stuck however this morning working on my latest features... Not sure exactly why... I have been drifting, searching for distractions...

and then suddenly... I am not even sure what did cause that but crypto did experience a mini crash... Volatility went through the roof and my system started to trade like crazy. It made something like 3000 trades in 5 minutes. It did trade a volume of over $100K in 5 minutes... It made about $400 of profits again in just 5 minutes... That is for that type of thrill that I have been working my ass off for now almost 3 years...

Overall things went very well. I think that I have only spotted 2 small glitches that I did fix and will release later when I upgrade my server...

$400 in 5 minutes!  Awesome!  Great work, Lano.