Spent two days upgrading and back testing model... Starting new 10 day trial...

[usma1992]

Feels good to watch winners roll in. I have had the best year I have ever had. With least amount of stress. How many people are still out there or did I lose a few?

I think my NBA picks are finally worth looking at... Started new trial today... I have spent at least 300 hours on the program... you would like to think ... I could get it to a profitable state.

Looking forward to Saturday's NCAAB picks...
Dave

[TheGoldenGoose]

AWESOME

[Yensel]

. Thank you, Dave!

[usma1992]

73-49 --- 59.8%

Solid... 2 Easy wins last night. Games weren't close... No bets today.

Dave

[usma1992]

JamesRav... You are correct...

Thank you
Dave
THE TEACH

[jamesrav]

Well... maybe I should just shut up and trust the system. UVA is an extremely unique statistical school Almost all their z-scores are below average. Their defense is extremely good... but they can't score which is rare. I looked at all the z-scores and thought... why would I ever bet the over on this game.

It predicted 72-70... score came in 72-68. CRAZY. I consulted all my UVA friends and we were all unsure why the system generated an over. I guess now we know. I will post picks later today. Good Night last night. 2-0 Spreads and Totals... 1-0 ML.

Dave

turned out to be an accurate prediction. Whenever you bring the word "sense" into the situation, you've gone from stat-driven to emotion, and that puts you in league with the 99% of bettors who go that route ... and lose in the long run. Unless you can scrape a ton of 'sentiment' info from social media, can test its usefulness and add it to the model if useful, it's always best to leave emotion out of it.

[usma1992]

Time	AWAY TEAM	PROJ AWAY SCORE	HOME TEAM	PROJ HOME SCORE		TOTAL	VEGAS SPREAD	VEGAS O/U		SPREAD BET	O/U BET
7:00 PM	SACRED HEART	79	STONEHILL	68		147	4.5	144.0		Away Minus F	Too Close
9:00 PM	UTRGV	59	GRAND CANYON	84		143	-22	154.0		Too Close	Under U

[usma1992]

Well... maybe I should just shut up and trust the system. UVA is an extremely unique statistical school Almost all their z-scores are below average. Their defense is extremely good... but they can't score which is rare. I looked at all the z-scores and thought... why would I ever bet the over on this game.

It predicted 72-70... score came in 72-68. CRAZY. I consulted all my UVA friends and we were all unsure why the system generated an over. I guess now we know. I will post picks later today. Good Night last night. 2-0 Spreads and Totals... 1-0 ML.

Dave

[jamesrav]

Don't bet UVA over... makes zero sense... will add small filter. The 142 projected should steer you away from an Under bet... but you should never bet this game over. Two hours to cancel bet.


Dave

why would that be? you have 142, well above the 127 along with probably the highest chance of a tie possible (with a 0 spread). Where does the "makes zero sense" come into play (or into the model). Are there heuristics involved that contradict the 142 total? I always remember that HAL in 2001 stands for Heuristics and Algorithmic.

[usma1992]

Don't bet UVA over... makes zero sense... will add small filter. The 142 projected should steer you away from an Under bet... but you should never bet this game over. Two hours to cancel bet.


Dave

[usma1992]

Games tonight... NBA had huge error that I just detected...

Time	AWAY TEAM	PROJ AWAY SCORE	HOME TEAM	PROJ HOME SCORE		TOTAL	VEGAS SPREAD	VEGAS O/U		SPREAD BET	O/U BET
6:30 PM	DETROIT	61	YOUNGSTOWN ST	89		150	-19	152.0		Parlay ML	Too Close
8:00 PM	RICHMOND	80	SAINT LOUIS	69		149	5	151.5		Away Minus F	Too Close
9:00 PM	VIRGINIA	72	BOSTON COLLEGE	70		142	0	127.0		Too Close	Over U

[usma1992]

Here is what I am considering doing and I am not sure how difficult it will be to accomplish... I think i am going to adjust the totals up and down and see how sensitive everything is and possible issue multiple types of bets. Not sure how easy it will be to do and how easy it will be to post. How many points should I add or subtract? I was thinking at the most two possessions up and down... so either 6 points or 4 points... down and and up.

In addition, I would do the same thing with the spread. Add and subtract from the spread and see how close I am to it becoming a bet, if it is a no bet.

I then can decide whether to buy points and whether it is worth it. It might become really complex... but might be worth it. For instance the most recent overs... I both lost by 2 points or less.... maybe it is worth buying two possessions. I am not sure whether it is or not.

Buying points might become too expensive.

Dave

[EasyCover]

Looks like picks are drying up in college. I have had a good season so I have decided to leave it alone and not bet until it tells me too. I am not opening it up any more. Whatever it is ... it is...

Dave
THE TEACH

Good policy. Only open it up if changes are made for tracking purposes.

[usma1992]

Looks like picks are drying up in college. I have had a good season so I have decided to leave it alone and not bet until it tells me too. I am not opening it up any more. Whatever it is ... it is...

Dave
THE TEACH

[usma1992]

I think I am lying to myself... but I might be close to breakeven or profitable on NBA. I am just trying to map college and pro together at some level. Many have told me to give up on NBA... but I have been told to give up on college before too.

I looked at two specific college games that I wasn't close on Saturday and then back tested. Often when I do that I find out that I mispredicted similar games with similar attributes. Either my system had an obvious weakness or the team match ups did.

I was able to eliminate two losing bets but I still would have lost and been 1-3. 1-5 on Saturday... 69-49... dropped to 58.4% with last Saturday.
My running total... that I quote doesn't include after the system adjustments. It only includes what the system predicted pre game. I think that's only fair.

When I post my overall numbers as a summary... the do take into the fixes and adjustments I have made.

Dave

[usma1992]

No bets tonight... my record is exceptional but I will have to wait until the next couple of seasons to see if it is legit. I am extremely happy with the product.

It came a long way this year for sure. Since Jan 2nd it has had a solid run. I opened it up Friday night just a little and it added one bet... Of course, it turned out to be a losing bet. Nonetheless, I will keep treking on.

Of course, NBA put me in my place again. I'm 0-2-1 on my newest trial. Both games lost yesterday.
6 FT missed in the last minute... of Montana St game. 18 overall as RM Logic stated. I confess, I didn't check the 18 number but I am sure he is right. 6 in the last minute.

Common man.
Dave

[Neek762]

jamesrav hit the nail on the head. I think I'm glad I went the blackjack route before getting heavy into sports betting because blackjack forces you to put emotion aside and trust the numbers and how things play out in the long run. But that only is true if you have the basic strategy down and know when to deviate. Goes hand in hand with sports betting. If you're a blackjack AP you're not just going to plop down at a 6:5 table and cross your fingers. If you're a sports bettor you're not going to open one sportsbook and pick your hometown team to win just because. Just have to use as many tools available and end of the day, still deal with variance lol. Literally no way around it unless you are fixing games or came from the future

[jamesrav]

at this stage, your model has either proven itself and stays steady at this elevated level (a tough task) or will slowly 'revert to the mean' which may or may not be profitable (hopefully the latter). Only time will tell. But tinkering with things isn't going to 'fix' a bad day. It's just probability doing its thing. When I'd lose 6 or 7 hands in a row at Blackjack, playing Basic Strategy (a break-even proposition), did I think Basic Strategy was 'broken'? Certainly not. Watching horse races the past several months for hours a day, I've seen more 30/1 , 40/1 and higher winners than I thought possible. And the Exchange payouts would have been even higher, maybe 60.0 up to 100.0. Weird things happen ; but an underlying 'good' model will always succeed over time. The Blackjack Card Counters are proof of that.

[Neek762]

I had 3 losses by a half point yesterday and one was also BYU

[EasyCover]

Gambling Gods... took me down yesterday... Actually showed no mercy... Maybe... by posting my record online... they decided it was time to teach me a lesson of humility. I went 1-5 yesterday and realistically was only in 3 or 4 games. Some games were complete blowouts that I never had a shot at.


As always, I will examine those games and see if I can come up with a better solution that works in the past and the future.

No games tonight... and I apologize if you lost significant money. I take my betting posts seriously and I don't want to let the public down. Either jamesrav or RM Logic or prolexus or someone else said respect the spreads... Apparently, I did not.

All the bets were unfiltered bets... so ultimately I have to decide before the end of the season whether I will include them after 2/3rds of the season is done.

I learn more from losing days... honestly, than winning days.

Two NBA bets are posted. Doing the best I can ... enjoy.

Dave
THE TEACH

Your record is awesome. Variance is always going to come into play. 60-62% is pretty much unsustainable. 55-58% is more realistic and what you should strive for. I did lose significant $$ yesterday on these but I pressed the buttons and took the risk. By the way, I have done well with the picks previously.

[RM Logic]

The picks werent that bad. Georgetown blew a 13 point lead because their coach is an idiot.
The Montana State game should have easily went over except there were 18 missed foul shots. Make just 2 of those and its a winner.
It should have been a 3-3 day.

I didnt play the BYU game until live in the 1st half and took BYU+8 and still lost.

[usma1992]

Gambling Gods... took me down yesterday... Actually showed no mercy... Maybe... by posting my record online... they decided it was time to teach me a lesson of humility. I went 1-5 yesterday and realistically was only in 3 or 4 games. Some games were complete blowouts that I never had a shot at.

As always, I will examine those games and see if I can come up with a better solution that works in the past and the future.

No games tonight... and I apologize if you lost significant money. I take my betting posts seriously and I don't want to let the public down. Either jamesrav or RM Logic or prolexus or someone else said respect the spreads... Apparently, I did not.

All the bets were unfiltered bets... so ultimately I have to decide before the end of the season whether I will include them after 2/3rds of the season is done.

I learn more from losing days... honestly, than winning days.

Two NBA bets are posted. Doing the best I can ... enjoy.

Dave
THE TEACH

[usma1992]

Ok... let me look into it... so basically throw out Under bets when the spread is within a certain window. Let me do my analysis ... I will let you know if it actually affects the results...

An Under bet may hit overtime more often than not if the spread differential is small.

Thank you for the insight. This is why I post ... in order to hear great opinions. Whether you are right or wrong... I will look into and it worth looking into.

Dave

[jamesrav]

First of all... thank you all for you questions or comments... they make me think and sometimes adjust or at least pay attention specifically on certain games to learn.

JamesRiv... I completely disregard OT regardless of the spread. In college basketball it happens 2% maybe 2.5% of the time. Moreover, how often do I have stated bets on that 2%. I feels more than it is. The other day I had Alabama Florida under... 174... they finished in regulation tied at 85-85. 170... The bet obviously lost in OT... but I record it as a win and only download regulation time scores. Once you start chasing anomalies your done. If OT happens and you win or lose who cares... percentage wise it doesn't happen that often.
Dave

certainly the tie situation is big in soccer and NHL, I can see why ignoring in basketball might be ok. But if your 2% figure is correct for all NCAA games, that doesn't then focus strictly on games with a 1 or 2 pt (or however high desired) spread. Maybe higher than 2%. In which case it might affect a decision.

When I was helping a guy formulate an NHL model, he had a bunch of 'anomalies' he wanted analyzed (such as if a team lost a home game badly, did they then play exceptionally better the next home game due to pride, etc). Whether anomalies are a distraction, or represent a better view of reality (and therefore a more 'finely tuned' model) is what makes modelling a major challenge. But you 'just' need to beat the Vegas opening line, when they are weakest.

[usma1992]

Time	AWAY TEAM	PROJ AWAY SCORE	HOME TEAM	PROJ HOME SCORE		TOTAL	VEGAS SPREAD	VEGAS O/U		SPREAD BET	O/U BET		Actual Away	Actual Home
2:00 PM	TEXAS ST	67	GEORGIA ST	77		144	-5.5	139.0		Too Close	Over F		0	0
2:00 PM	BYU	78	KANSAS ST	67		146	1	147.0		Away Minus U	Too Close		0	0
6:00 PM	GEORGETOWN	81	DEPAUL	68		149	4.50	151.0		Away Minus U	Too Close		0	0
8:00 PM	SACRAMENTO ST	68	MONTANA ST	76		144	-7.50	131.0		Too Close	Over U		0	0
9:00 PM	CS-NORTHRIDGE	70	CS-FULLERTON	66		136	-1.00	147.0		Too Close	Under U		0	0
1:00 PM	ALBANY	71	UMASS-LOWELL	77		148	-8.50	163.5		Too Close	Under U		0	0

Added one game...
Dave

[usma1992]

Jan 26-Present Filtered and Unfiltered

-4.0	PAR ML	15.0	10	OVER	0	9	0.5	UNDER	5.5	6	10		5	-12		11.00	-9		1	4.5
150	8	14	3	-2.6	7	0	4.25	-3	-6	AWAY -	0.7		HOME -	-6	-3.0	AWAY+		2.0	HOME+	-2.75
8.0	8.0	100.00%	23	14	60.87%	16	13	81.25%		17	11	64.71%	14	8	57.14%	2.00	2	100.00%	3	3	100.00%	75.0	51.0	68.00%

The far right statistic DOESN'T include ML bets. I track them but their payout is extremely low.

Not sure whether u can decipher my analysis but I did the best I can.

All of the information is current. I worked to create algorithms over the course of the first half of the season. Then since Jan 2... I have run this test. When Jan 25th approximately 2/3rd of the season complete... I decided to remove filters and see what my predicted model would do without filters. Jan 26-Present include filter and non filtered bets.

Dave

[usma1992]

Nov 18-Jan 2 Filtered and Backtested

-4.0	PAR ML	15.0	10	OVER	-1	9	0.5	UNDER	5.5	6	10		5	-12		11.00	-9		1	4.5
150	8	14	3	-2.6	7	0	4.25	-3	-6	AWAY -	0.7		HOME -	-6	-3.0	AWAY+		10.7	HOME+	-2.75
62.0	61.0	98.39%	81	53	65.43%	30	25	83.33%		17	16	94.12%	40	30	75.00%	36.00	24	66.67%	24	19	79.17%	228.0	167.0	73.25%

Jan 2- Jan 25 Filtered

-4.0	PAR ML	15.0	10	OVER	-1	9	0.5	UNDER	5.5	6	10		5	-12		11.00	-9		1	4.5
150	8	14	3	-2.6	7	0	4.25	-3	-6	AWAY -	0.7		HOME -	-6	-3.0	AWAY+		0.3	HOME+	-2.75
9.0	8.0	88.89%	16	12	75.00%	4	4	100.00%		6	6	100.00%	15	10	66.67%	2.00	1	50.00%	2	2	100.00%	45.0	35.0	77.78%

Not sure whether u can decipher my analysis but I did the best I can. Far right does not include ML bets.

All of the information is current. I worked to create algorithms over the course of the first half of the season. Then since Jan 2... I have run this test. When Jan 25th approximately 2/3rd of the season complete... I decided to remove filters and see what my predicted model would do without filters. Jan 26-Present include filter and non filtered bets.

Dave

[Neek762]

Do you check odds screens at all throughout the day, or just use what the model shows is the best bet for the day? Just asking since you bring up the UConn - Creighton game. There was some info on the game combined with line movement that made me take Creighton. There were some aspects on KenPom where Creighton was better, and adding that in with a true road game for UConn seemed to work out.

[usma1992]

First of all... thank you all for you questions or comments... they make me think and sometimes adjust or at least pay attention specifically on certain games to learn.

JamesRiv... I completely disregard OT regardless of the spread. In college basketball it happens 2% maybe 2.5% of the time. Moreover, how often do I have stated bets on that 2%. I feels more than it is. The other day I had Alabama Florida under... 174... they finished in regulation tied at 85-85. 170... The bet obviously lost in OT... but I record it as a win and only download regulation time scores. Once you start chasing anomalies your done. If OT happens and you win or lose who cares... percentage wise it doesn't happen that often.

RM Logic... my models are straight math driven... which have their positives and negatives. I am not looking at current trends... but obviously the data reflects outcomes etc... You indicated that Ranked vs. Unranked were 1-8 the last week. How many of those games did my system pick? UVA-VTECH no... CREIGHTON--UCONN... no. I think you are more susceptible to variation anomalies if you are playing trends. Stay away stick to basic parameters on what games work... what games don't... ranked unranked who cares...

Then again... you have to decide to bet on BYU or not. I am not a huge fan of the bet... but I am now completely convinced the model will make me a winner and I toss out my person bias. Thats where I sit.

My thoughts
Dave

[RM Logic]

Thanks for the early Saturday picks.
Ranked road favorites at an unranked home dog have been getting killed lately. Last week they were 1-8. And the week before they were 5-8 I think. Ranked BYU as a favorite at Kansas State makes me pause of course as the home dogs have been unreal lately. Any thoughts on this trend?

[jamesrav]

for the Davis vs. Northridge game - expected to be close (you have a 2 pt differential) - does your model take into account the possibility of a tie in regulation, which would affect the O/U outcome? the chance for a tie seems much higher than if the spread was 10 or something. Or are ties in regulation quite rare and get swamped by the normal outcomes?

(I now see it ended 66 65, so it flirted with the tie situation )

[usma1992]

Neek I appreciate the feedback... I will be more transparent... in my follow on posts...

Dave

[usma1992]

Saturday's games...

Time	AWAY TEAM	PROJ AWAY SCORE	HOME TEAM	PROJ HOME SCORE		TOTAL	VEGAS SPREAD	VEGAS O/U		SPREAD BET	O/U BET
2:00 PM	BYU	78	KANSAS ST	67		146	1	147.0		Away Minus U	Too Close
6:00 PM	GEORGETOWN	81	DEPAUL	68		149	4.50	151.0		Away Minus U	Too Close
8:00 PM	SACRAMENTO ST	68	MONTANA ST	76		144	-7.50	131.0		Too Close	Over U
9:00 PM	CS-NORTHRIDGE	70	CS-FULLERTON	66		136	-1.00	147.0		Too Close	Under U
1:00 PM	ALBANY	71	UMASS-LOWELL	77		148	-8.50	163.5		Too Close	Under U

[Neek762]

But what I'm getting at is, as I mentioned before, keeping a running record of all wagers would help differentiate which wins and losses are sides - totals - MLs. Unless you just want everyone on this thread to keep their own running count. ROI is also way different than WL Record. 68-44 with heavy favs is different than 68-44 ATS, or if someone's middling there can be a lot of skewed win loss records but ROI will reflect the success. Maybe make a Pikkit or BetStamp account we can follow

[usma1992]

I post my picks daily... as I'm sure you know... you can also go back and look.

Looks like 5 Totals no spreads.

Dave