MGCOLBY's Picker Thread 2024

[mgcolby]

Ok sorry for not updating this yesterday. The excel file became a total shitshow. I spent every free minute I had yesterday and my entire evening unfucking the shitshow that it became. I have still have to work out the automation of the lines. So be careful and double check those, which you should be anyway as the lines are constantly moving. The prediction data is accurate and correct IAW the formula. Here is today's game. I included the scores of the games already played.

Edit Note: the yellow highlighted text indicates a predicted win for the dog based on this morning's lines.

 

[mgcolby]

Can A.I. automate it for you?

That is what I am using to write the vba macros. I am pretty solid with M query for Power Queries in excel. But this is pretty complex from just from an organization perspective. And ironically that is what led to my shitshow yesterday. Just blindly following the Grok code changes which work pretty well in general. But it does cause issues when working on different vba scripts within the vba project. Essentially after so long it forgets the base parameters of various functions and modules and how they work together. When that happens it inserts calls within the script code that are already a shared function for all the modules. This caused a bunch of corruption. Then I went back into the versioning and opened a previous version, but then I needed to work back to where I was etc...it sucked. It is now functioning again with a bit of manual labor. I will work on that this weekend.

But to your overall question, yes! I wouldn't be as far on this as I am without it. This has also opened my eyes for ways to streamline the NFL picker. But more importantly, it has shown me that I can essentially create one for any sport assuming I can come up with a formula that works for that sport. But that is down the road.

 

[mgcolby]

Here is yesterday's results. I am running the script to scrap the data and it takes a bit hour to hour and a half. Its been running for about 20 minutes. Hopefully I will have it up in time for the 3pm games. There are 100 games today.

 

[mgcolby]

A little late but here it is:

Still a lot of manual entries on the lines. Hence why the Moneyline column is not completely filled out. Ran out of time. Lines are constantly moving so as always verify the lines are still accurate when placing the bet. Yellow are dogs it predicts to win outright.






And the results. Its first losing day on the totals. Split the games on the spread and still doing really well picking games outright. Will have today's games posted soon. Hopefully before 11:30.

 

[mgcolby]

Here is todays games. A little later than I wanted.



And the daily results from yesterday 03-09-25

 

[mgcolby]

Ok worked on some of the aesthetics yesterday. I'm sure I will tweek it some more. But I think the overall color scheme is easier on the eyes and easier to read. I changed up a few things with the color scheme. I also included a legend to explain it.

1. I color coded the totals where there is a 10+ differential. For the slower among us i.e over differential 10.3 and under -10.3 will be color coded green. I thought about color coding the groupings i.e. 6-9.9, 3-5.9, under 3 etc...not sure if that would be color overload or not.
2. As mentioned, there is a legend in the bottom left
3. Not sure if I like the different shading of the columns for the records. Open to feedback on that.
4. Will work on record keeping this week. Especially on the differentials.
5. Let me know if you like the new format. Always open to suggestions and feedback.




Went on an epic 7-0 run and made a rookie mistake of texting my son about it. I toyed with angering the gambling gods and got appropriately godsmacked for my sins! I proceeded to lose the next game that ended (Charleston NC Wimington) in a crazy fashion. Then finished up the night going 1-3. For a total of 8-4. I played all the totals blind. Overall the picker had a good day. Won both dogs, which I avoided betting on because I got my ass kicked by them yesterday. And Delaware was a +240 at tipoff. Speaking of odds changing, I did update the spreadsheet to reflect the totals at the time of my bets. which were just a few minutes after posting the plays this morning. I have discovered lines move fast and sometimes drastically in college basketball. Anyway they are highlighted in yellow. I included the changes in the legend. Full disclosure and all.

 

[mgcolby]

Want to add a few notes. And yes, this is about the time I have been working on this shit to at night. I hope you fuckers appreciate it. Just kidding I would do it anyway. Because I use it. I have been dogging Grok making that fucker burn some serious CPU coding this thing for me. It has been a life saver. No way I would be as far along with this as I am or even sharing it at all. It has automated so many parts of this workbook for me, it's been unbelievable.

Tonight:
1. I was finally able to get the task scheduler working, which I use all the time for work. But I couldn't get it to run the damn script to pull the latest game data. But we (Grok and I) persevered and got it figured out. Python installation locations is a finicky thing.
2. Created a script to automate the importation of the updated data into the data table in the main excel workbook. This is a big timesaver as manually doing it was a pain in the ass because of the way excel handles copy and pasting.
3. I was able to tidy up some of the aesthetics as you guys saw. Hope you like the new layout. Another win I believe for Grok as it suggested the color scheme improvements based on my base colors I seem to like.

Hopefully tomorrow night I will be able to automate this thing about 99.9% of the way. So, when I wake up in the morning the picker is sitting pretty waiting for me. Once that is complete, I have a few things I need to work on with some of the normalizations of the names between the various sites I pull the data from i.e. game data, schedule etc... that will be a work in progress for a little bit. But the other day it auto filled 96 out of 100 games with the various lines which was a huge fucking improvement. Then I guess at some point I will set up a web scraper to gather up the scores and populate those.

I am excited for things I have learned with this, because it is going to improve the overall process for updating the NFL picker this year. Which will already be easier as I will not be tracking three different versions.

Anyway, I have rambled on enough.

 

[Kontradiction]

Want to add a few notes. And yes, this is about the time I have been working on this shit to at night. I hope you fuckers appreciate it. Just kidding I would do it anyway. Because I use it. I have been dogging Grok making that fucker burn some serious CPU coding this thing for me. It has been a life saver. No way I would be as far along with this as I am or even sharing it at all. It has automated so many parts of this workbook for me, it's been unbelievable.

Tonight:
1. I was finally able to get the task scheduler working, which I use all the time for work. But I couldn't get it to run the damn script to pull the latest game data. But we (Grok and I) persevered and got it figured out. Python installation locations is a finicky thing.
2. Created a script to automate the importation of the updated data into the data table in the main excel workbook. This is a big timesaver as manually doing it was a pain in the ass because of the way excel handles copy and pasting.
3. I was able to tidy up some of the aesthetics as you guys saw. Hope you like the new layout. Another win I believe for Grok as it suggested the color scheme improvements based on my base colors I seem to like.

Hopefully tomorrow night I will be able to automate this thing about 99.9% of the way. So, when I wake up in the morning the picker is sitting pretty waiting for me. Once that is complete, I have a few things I need to work on with some of the normalizations of the names between the various sites I pull the data from i.e. game data, schedule etc... that will be a work in progress for a little bit. But the other day it auto filled 96 out of 100 games with the various lines which was a huge fucking improvement. Then I guess at some point I will set up a web scraper to gather up the scores and populate those.

I am excited for things I have learned with this, because it is going to improve the overall process for updating the NFL picker this year. Which will already be easier as I will not be tracking three different versions.

Anyway, I have rambled on enough.

Yeah that’s a shitload of work. Grok is awesome. It’s actually helped Guts design our new DingleKings sports book.

 

[mgcolby]

Tuesday 03-11-25

Look at me getting these out nearly an hour before the first tip-off. The automation helps a lot. Had a couple of hiccups that need some working out. And today's picks. To hopefully a winning day for all!




Yesterday's results: Not great but a winning day across the board, which is always a good day.

 

[PM15]

@everyone I know many of you are into March Madness. I'm not...because I'm not gay.

But you all might want to keep an eye on this heading into the tournament. Great stuff, @mgcolby

Going to start to fill my brackets soon.

 

[mgcolby]

Going to start to fill my brackets soon.

Hopefully this thing works well for it. Its really liking the overs in these conference tournaments which I don't like too much especially when teams play back to back nights.

 

[mgcolby]

03/12/25

Got it in a little late today. That work stuff gets in the way. Here you go today's picks enjoy!




View: https://youtu.be/gH476CxJxfg

 

[mgcolby]

3/13/25

In with 13 minutes today!



And the results:

 

[mgcolby]

Had some technical difficulties today. My laptop rebooted in the middle of running the script and I didn't realize it until now. So running it again and it takes a couple of hours.

 

[mgcolby]

Here is the basic record. I will at some point break down the differentials. Not sure when I will get the time to do it.

It's not great but its not "losing" in terms of games.

​

 

[mgcolby]

03/14/25

As promised.



Results

 

[mgcolby]

03/15/25

Record

​

Results

​

 

[mgcolby]

03/16/25

Record

​

Today's plays.

 

[mgcolby]

03-18-25

Ok, we have some substantial changes to the picker. While all you lushes were out striking out on scoring with Shamlock in Southie, you fucks even missed out the whales from what I have been told! What a pathetic bunch as Shamlock would say the whole lot of ya! Any way, while you fucks weren't getting your peckers wet, drowning your sorrows away, drinking fucking schlitz and trying to sing Piano Man in an Irish accent! I was in the lab slaving away crunching numbers and shit!

This fifty-fifty shit doesn't sit well with me. I needed a way to improve this thing before it goes down quicker than Dingle when Roseanne spreads her legs and parts her bush with a rake! So I figured why not use AI. I won't get into the particulars as the colonial can not give away the secret recipe. I did manage to come up with a confidence rating model that when run through the previous weeks improved upon its current scores.

This is a 1-10 scale with the baseline being a 5. Doubt you'll see a pick be below or much below a 5. Haven't seen one yet given the short life span of this picker. I will work on the presentation as I move foward but for now will leave it as is with the confidence score columns added. I also worked with AI, based on the confidence rating model, to come up with 3-5 picks of the day along with a 3 leg parlay. The three leg parlay will not neccasarily be the top 3 plays of the day. We have rules built in to the picks of the day. They can't be excessive favorites on the ML say a +1800 for example. We cap those off at -300 and weigh risk vs reward when factoring that in. For the parlays we do remove the cap but the parlay has to have realistic value again with risk vs reward in mind.

EDIT: Adding the win-loss table for the historical data. Its not 100% of the data for full disclosure: AI was having some issues processing a worksheet or two. It represents about 80% of the games during the picker's infancy.

Confidence	Total W	Total L	PCT%	Spread W	Spread L	PCT%	Moneyline W	Moneyline L	PCT%
5	132	140	48.5%	119	121	49.6%	169	133	56.0%
6	31	32	49.2%	33	30	52.4%	33	21	61.1%
7	39	17	69.6%	36	20	64.3%	25	12	67.6%
8	11	4	73.3%	10	5	66.7%	4	1	80.0%
9	0	0		0	0		0	0

I would say tread lightly but hopefully this homes in on the better plays.

Picks of the Day:
St Francis PA vs. Alabama St: Under 138.5
Wichita St +6.5
CS Northridge +7.5

Parlay of the Day:
St Francis PA vs. Alabama St: Under 138.5
Chattanooga: ML
CS Northridge +7.5
+681

 

[mgcolby]

03-19-25

Here is today's picks. I will put in the top AI picks based on the model. And I will add those as an edit shortly.


Top 3 Plays
AU/Dayton - Under 153.5: Confidence 7.9
Xavier/Texas - Over 151.5: Confidence 7.9
Loyola Chi: MoneyLine Loyola Chi: Confidence 7.8

3 Leg Parlay
FAU/Dayton: Under 153.5 (-112).
Xavier/Texas: MoneyLine (+136).
Loyola Chi: MoneyLine (-166).
Odds: +484.