TheKingfish

@boro33

You are welcome. And to MrFreedo, I offer some trailing commentary to assuage the confusion. First, the context. Most every ‘technical’ handicapper is in the business of crafting trends from-history that will be applicable to the future. So, as a Power-Rating (PR) capper, I share a classic-research stumbling-block example:

Let’s say I have this researched trend: Bet a Home Fav with a PR >= 89. It has reasonable volume and a success rate of .575, or some facsimile.

Structurally, that’s a trend with a binary component (category) and a variable component (PR). Think of Line as the x-axis, and PR as the y-axis. However, in this case, that fluid Line is divided into two categories.

This trend has an application problem due to that binary aspect; the rigid toggle between dog & fav.

If the Line moves thru zero (-1 >> +1), our coveted trend evaporates. So, on the field, even though that new Home Dog is not substantially different from the old HF -- the handicapper must abandon it. S/he has no data to support the HD it has become.

Mentioned prior, another large/inherent problem with this trend structure is line magnitude. Is that -1 HF identical to a -10 HF? Of course not; there are differing performance sectors along that x-axis. However, here’s the larger analysis problem -- optimum-seeking research employing TWO (2) variables is a challenge.

That most-optimum trend is difficult to identify. Successful research might state it as follows: Bet a Home Team having a PR >= 89, and a Line between -2.5 and +6.5.

Most-optimum being the operational phrase here. Tech stuff tends to be user-centric and often defies explanation -- but I hope that helped.
Good luck,
TheKingfish

@IntenseOperator

Thanks for your posted support -- much appreciated.

My content tends to be an acquired-taste, as it was described by a (TalkSport) forum friend of long ago. Thus, typically, I get few players who wade-in with a reply. However, post algorithm adjustment, it was gratifying for me to bring some value to the forum during the final eight weeks of 2024 (top eleven, .650, 52-28-3).

Suffice to say, my madness is multi-faceted. And with research now pulling me back into the dungeon, I will happily see you all again in August......
Good luck,
TheKingfish

@MrFreedo

Thanks for your reply. Clarification is always a good thing. I will answer your queries in opposite order.

My large historical database purposely does not contain neutral-site games. Therefore, the algorithm has no capability to evaluate them. And if those games were logged and processed separately -- that sample size would be too small to yield much predictive confidence. Leastwise, at a level I am comfortable with....

No systemic filtering to distinguish dogs and favorites. Categories are problematic, in that moving thru-zero changes the dynamic. For example, isn’t a 1-point favorite nearly identical to a 1-point dog? And to step further, is a 3-point dog the same as a 14-point dog? Instead, my system avoids that kind of rote category-lumping by focusing on the seamless combo of home-team and line magnitude.

Kindly reverse your perception. In non-robot terms, I am saying if you bet the line early you have more value. When The Man demonstrates a lower AveSM (late week), he is theoretically achieving more predictive-strength – to the disadvantage of players.

That said, we all know the line is not intended to be a predictor – but rather, a bettor-perception value intended to equalize intake ($). Questions and comments welcome.
Good luck,
TheKingfish

A short one, so as not to cloud the mind with onerous Tech -- in the midst of the final CFP games. SpreadMargin (SM) is the amount by which a team covers, or fails to cover. Given that my regular season database only recently finalized itself (2013-24, 7583 games), I thought I’d share an offshoot. Specifically, I’ve done some Average SM work prior to the launch of my off-season research:

12.292 2013-23
12.151 2024
Comparatively, this means The Man was a better-predictor in season 2024 -- by 1.15 percent ([12.292-12.151]/12.292). Not good news.

The above is final line data.
12.278 2013-24 final line
12.396 2013-24 early line
Comparatively, this means The Man has been a 1.03 percent weaker-predictor when players bet the early line (12.396/12.278). A slight player edge......when s/he bets into a line that is less mature/strong.

So, what might we infer from this, if anything? The second piece constitutes a capper strategy. A small thing that maybe slides into the back pocket.....

If valid, the first piece would seem to have more import. Assuming we have Artificial Intelligence operating, and The Man grows one percent better every year – we’re in trouble. However, given the Law of Diminishing Returns, it’s unlikely such an improvement magnitude could be sustained. However again, the Law of DR is also subject to our new world of AI.

So then. Is it a case of having fire where there is smoke? Or, OTOH, maybe I’m just “blowing” smoke......
Good luck,
TheKingfish

Hey now. Always good to connect with another refugee from SportsTalk!

Despite the above talk about vaunted "system competitiveness” my Hilltoppers were far from that last night. Indeed. A technical anomaly no doubt.....

Ah but, ever the degenerate gambler, I stumbled upon the fact that there is a fourth non-neutral site game this week. I suspect it’s a late add (????). Duly capped, the wallet flips open one more time:
Marshall(+6)/LA-Lafayette

Noticed off hand – the Herd actually has a phenomenal ATS record this year. I only bet non-advantageous line magnitudes (like a +6 dog) -- when I’m forced to. So, with this 7:30 game, color me a ‘patient’ degenerate gambler. For the record though, the box likes it at any magnitude.
Good luck,
TheKingfish

A winner was posted here last Saturday (10-8-2, .556), capping a string of eight (8) straight winning weeks. The full-season performance finishes at .553 (94-76-5); a unique combination of both volume & success.

Let’s visit rank-value during those eight (8) weeks since the algorithm was adjusted. Over that time, the top eleven best-bets have gone .650 (52-28-3). Stellar, and a distinct double-down on the volume/success combo mentioned above.

Alas, it wouldn’t be a bona-fide Kingfish post without some (boring) tech talk. Was there a drill-down silver lining in the last week 10-8-2 performance that may bode well for the future? Maybe. Consider SpreadMargin (SM); the amount a team covers by, or fails to cover by.

Beyond wins & losses, the algorithm’s competitiveness can be secondarily judged by SM. The 6844-game database average SM is 12.3. So, theoretically, if your teams are winning by more than that, or losing by less than that -- you're doing well. You are competitive and, invariably, you will win your share of games. Especially if you have that predictive quality over the large volume of games.

Here's a small-sample last-week example. Of my 10 non-covering best-bets -- the SM average was 8.95. That was 73% (8.95/12.3) of the long-term SM average, and just 2 of those 10 no-cover best-bet outcomes exceeded the SM average. Both of those observations point toward positive structural performance. Humble opinion.....

Lest I forget this week! We have the antithesis of large opportunistic board volume – the conference championship games (8). My research database holds no neutral site games, and with just 3 of those 8 on home turf – my capping is limited. Thus, the box offers only the Hilltoppers as bankable:
WesternKentucky(+5)/JacksonvilleSt

It has been enjoyable to post in 2024, and I hope this corner brought you some value/entertainment. Until next year (if/when I return), I wish you the best.
Good luck,
TheKingfish

For the record.......an operator error to report. A certain element of the algorithm is brand new – and I totally misread it earlier today. Apology as needed, but I’m actually flipping my pick on that last game.
6.5) AirForce(-5)/SanDiegoSt

With that reduced line, I have not placed yet but, as it turns out, the box likes the Falcons up to and beyond -7.
Good luck,
TheKingfish

Monitoring them all, this is a late-addition based on line movement:
6.5) SanDiegoSt(+5.5)/AirForce

Though I have not actually placed this one yet.....

Given that the Aztecs host the last game on the board tonight, the HA is figuring we might have those numerous favorite-lovers push the number upward. Please note in passing -- I am allowed to make “timing” decisions. However, in the end (FYI), the algorithm’s basement is the current 5.5.
Good luck,
TheKingfish

Success visited this corner of the forum last week (9-7, .563). The full-season performance record is now .553 (84-68-3). That unique combination of both volume & success is an accomplishment (humble opinion), so no complaints. With that, the box shows all of these on this week’s ticker tape:
1) AppState(+2.5)/GeorgiaSouth
2) LA-Lafayette(-9.5)/LA-Monroe
3) NorthTexas(-11)/Temple
4) Toledo(-8)/Akron             L
5) USC(+7.5)/NotreDame
6) MiamiOhio(+3)/BowlingGreen
7) UAB(+3)/Charlotte
8) Marshall(+3.5)/JamesMadison
9) JacksonvilleSt(+2)/WestKentucky
10) Stanford(+3)/SanJoseSt
11) UTSA(+7.5)/Army
12) Pittsburgh(+5)/BostonCollege
13) KansasState(+3)/IowaState
14) NorthCarSt(+3.5)/NorthCarolina
15) WestVirginia(+3.5)/TexasTech
16) Virginia(+7.5)/VirginiaTech
17) Navy(+2.5)/EastCarolina
18) UtahState(+6)/ColoradoSt
19) CoastalCar(+1)/GeorgiaSt

System rank value had taken somewhat of a backseat since the algorithm was formally adjusted seven weeks ago. However, ever the data-watcher, the system’s upper-strength has asserted itself. Over that time, the top six have gone .683 (28-13-1).

Some requested slack on the above. For the record, I have yet to place on several best-bets due to their ‘intractable’ short-week non-advantageous line magnitudes (+2.5, +6, -11). I won’t get what I want on most – but like most of my ilk (degenerate gamblers); I’m an optimist.

An observation. The algorithm normally stays pretty low key; rarely venturing out to predict marque games. Justified or not, the Human Administrator has always liked that subtle characteristic, feeling like big-money contests also produce Big Scrutiny -- and ultimately a tougher line. However, the box was on both of last Saturday’s large SEC upsets (Okla/Alabama & Florida/Miss). Further, Florida opened at +6.5 and closed at +13. So, among our wagering brethren, the Gators were very much a contrarian pick. As they say, a dog that barked.......

This is our last major-volume week. And a first for season 2024 -- we have the full-slate of 67 games. Here’s hoping you can turn that volume-opportunity into dollars, along with a grand experience today. Happy Thanksgiving!
Good luck,
TheKingfish

For the record:
Toledo(-8)/Akron

With the board at full-volume this week, the usually-prolific algorithm is looking to embrace 15-18 best-bets. For what it’s worth, it will have the Rockets somewhere in the top four (4). That rank will post Thursday morning -- a prelude to your afternoon turkey!
Good luck,
TheKingfish

Some small success was divined here last week (6-5, .545). That mostly holds with the full-season system performance (75-61-3, .551). The box knows only incoming data, and has not a single-shred of bad-beat prejudice. Nor does it carry the opposite burden of cherished teams. This week, in typical rote fashion, the box likes all of these:
1) Buffalo(+2.5)/EasternMich        W
2) Rice(-6.5)/UAB
3) OhioUniv (+2.5)/Toledo        W
4) Oklahoma(+14)/Alabama
5) Florida(+10.5)/Mississippi
6) AirForce(+4)/Nevada
7) LA-Monroe(+3.5)/ArkansasSt
8) ColoradoSt(+3)/FresnoState
9) EastCarolina(+3)/NorthTexas
10) NewMexicoSt(+4)/MiddleTenn
11) SanDiegoSt(+4)/UtahState
12) WestKentucky(+1)/Liberty
13) Marshall (+2.5)/OldDominion
14) GeorgiaSouth(+2.5)/CoastalCar
15) SamHouston(+ 6)/JacksonvilleSt
16) NoIllinois(+3)/MiamiOhio             L

Save for just one; all dogs. Rank value has taken a backseat since the algorithm was adjusted six weeks ago. Performance during that span has been .652 (43-23-1).

I know tech-talk is a hard sell on the forum, generating few replies, but I wonder if anyone employs an early-week quantitative handle on point spread magnitude. As in -- when to strategically hold, and strategically place (???). For example, the most compelling dog-hold level would be +2.5. And conversely, a timely persuasive favorite-bet would be -6 or -6.5. The power aspect of this logic feathers out with larger line magnitudes, but such strategy can be dollar-return significant over the course of a full season. Most pronounced of course, when the player has volume ability.

Our concluding two wager-weeks in 2024 will be bursting with opportunity. This week we have no less than 61 on the board. And next week (Thanksgiving) will have all 134 FBS teams in action; a full slate of 67 games. That will be a veritable “opportunity bonanza” for techies able to evaluate each and every contest!
Good luck,
TheKingfish

For the record:
NoIllinois(+3)/MiamiOhio
Buffalo(+2.5)/EasternMich
Ohio(+2.5)/Toledo

Said earlier in the year, the 2024 algorithm tends to embrace teams that promise to be SU competitive (low spread). I recognize and like that MO, although research (and not the Human Administrator) controls that.

Good luck,
TheKingfish

Just performing my civic duty with an added post. Has anyone been watching the Sam Houston line? The Bearkats are at -19.5 in some places. And if you got them early (-13.5 open), then you’re sitting on an attractive gap (6). Plus.......there could be a bit more upside before game time.

Of course, that second-source ‘middle’ play would be Kennesaw St. Middles at that line magnitude are less likely than somewhere closer to zero, but it’s still thoughtfully tantalizing. The game goes at 3:00 EST.
Good luck,
TheKingfish

Some bona fide success here last week (8-3-1, .727). That almost pushes the full-season large-volume system performance into respectability (69-56-3, .552). Turning the proverbial crank, the box has all of these in the win column this week:
1) BoiseState(-13.5)/SanJoseSt
2) SamHouston(-14)/KennesawSt
3) Liberty(-13.5)/Mass
4) WashingtonSt(-13)/NewMexico
5) JamesMadison(-3)/OldDominion
6) Troy(+8)/GeorgiaSouth
7) Rutgers(+6)/Maryland
8) UCLA(+3.5)/Washington
9) Georgia(-9)/Tennessee
10) MichiganState(+3.5)/Illinois
11) ArkansasSt(+3)/GeorgiaSt

Rank value has almost become a moot topic since the algorithm was tweaked five weeks ago. Performance during that span has been an admirable .673 (37-18-1).

I cap early Monday, and casually do likewise every morning during the week. I also place some on Monday; then as I see fit afterward. Opportunity is the goal of course, in addition to an emerging ‘middle’ that might later drop fortuitously into my lap. Related, and purely an anecdotal observation in 2024 – daily line movement has been rather stable; not large. If statistically true, that would not be good news for cappers; suggesting growth in The Man’s strength & confidence. Perhaps due to that now-ubiquitous enhancement known as AI.

53 on the board this week; up from 50 last week. Of course, any boost in opportunity makes us (examine-all) technical cappers happy. Albeit, it seems there are not many of that genre on the forum......
Good luck,
TheKingfish

Some success here last week (5-4, .556). Just barely over breakeven for the full season (61-53-2, .535). But, not so good on the volume front. The box likes these this week:
1) Army(-3)/NorthTexas
2) Connecticut(-6.5)/UAB
3) Liberty(-10.5)/MiddleTenn
4) CentralFlorida(+4)/ArizonaSt
5) BowlingGreen(-12.5)/CentralMich       L
6) LSU(+3)/Alabama
7) NewMexico(+3)/SanDiegoSt
8) SanJoseSt (+3.5)/OregonState
9) Duke(+3)/NorthCarSt
10) Rice(+9.5)/Memphis
11) Virginia(+7.5)/Pittsburgh
12) AppSt(+ 1)/CoastalCar          L

Since I adjusted the system four weeks ago, the top six have been strong (17-7, .708).

With ten weeks gone and four to play, are we interested in some tech-talk on category performance? Well, thus far the Home Dogs are .500 (86-86-4), and the Away Dogs are .517 (163-152-9). ATS macro data like that runs to .500, so no surprises. And no profit in wagering one or the other......

A bit more (boring) tech to share. I bet both early and late, so am always in the hunt for a middle opportunity. Of course, you need to be on the right side of that gambit early. No small thing! My personal line-movement target is >=4. So, how often has that occurred in 2024? That level of early/late movement has happened exactly 26 times (5.2%).

50 on the board this week; up from 49 last week. More opportunity -- always a good thing.
Good luck,
TheKingfish

For the record:
BowlingGreen(-12.5)/CentralMichigan
AppState(+1)/CoastalCarolina
NewMexico(+3)/SanDiegoSt

Good luck,
TheKingfish

Success last week (10-4, .714). Am just barely over breakeven for the full season though (56-49-2, .533). Not good. The venerable box likes these this week:
1) Tulane(-14.5)/Charlotte       W
2) Arkansas(+7.5)/ Mississippi
3) JacksonvilleSt(+2.5)/Liberty      W
4) MichiganSt(+8)/Indiana
5) UTSA(+7 .5)/Memphis
6) Rice(+12.5)/Navy
7) CentralFlorida(-6)/Arizona
8) SoAlabama(-6)/GeorgiaSouth
9) TexasSt(-4)/LA-Lafayette        L

Since I tweaked the system (three weeks ago), the top six has performed well (13-5, .722). With five weeks left, there’s a glimmer of hope for me……

49 on the board this week. Down from 56 last week; thus less opportunity.
Good luck,
TheKingfish

Forgot to share this one earlier.
Tulane(-14.5)/Charlotte

Good luck,
TheKingfish

For the record:
TexasSt(-4)/LA-Lafayette
JacksonvilleSt(+2.5)/Liberty

Good luck,
TheKingfish

Some success here last week (8-5, .615). For the full season it’s 46-45-2 (.505). Not good. Data into the box; give the handle a turn; the outcome:
1) TexasA&M(-1)/LSU
2) GeorgiaTech(+10)/VirginiaTech
3) Auburn(+2.5)/Kentucky
4) ColoradoSt(-6)/NewMexico
5) FresnoState(-6)/SanJoseSt
6) Virginia(-4)/NorthCarolina
7) TCU(-5.5)/TexasTech
8) Ohio(-6)/Buffalo
9) Connecticut(-6)/ Rice
10) Minnesota(-4)/Maryland
11) BallState(+12)/NoIllinois
12) NorthTexas(+7.5)/Tulane
13) BostonCollege(+7.5)/Louisville
14) Duke(+13)/SMU

Some rank-value was had last week -- five of the top seven were winners. Eight weeks gone; six weeks left.

56 on the board this week. Down from 59 last week, so the opportunity quotient has declined a bit.
Good luck,
TheKingfish