TheKingfish

For the record – adding two more.
5.3) Clemson(-5.5)/SMU
5.6) SouthFlorida(-21)/FloridaAtlantic

Good luck,
TheKingfish

A winner here last Saturday (8-5, .615). In fact, the top eight went 6-2 (.750). So far, five winning weeks out of seven. Albeit, producing an overall of just .528 (47-42), which is barely above breakeven (.524). Assessing every game on the board, the box likes all of these:
1) CentralMich(+6.5)/BowlingGreen
2) SanJose(+4.5)/UtahSt
3) Connecticut(+2.5)/BostonCollege
4) Michigan(-5.5)/Washington
5) SamHouston(+3)/UTEP         L
6) Georgia(-6.5)/Mississippi
7) Iowa(-2.5)/PennState
8) Alabama(-8)/Tenn
9) Vanderbilt(-2.5)/LSU
10) NotreDame(-8)/USC
11) Rutgers(+17.5)/Oregon
12) Virginia(-17.5)/WashingtonSt
13) California(-10)/NorthCarolina
14) Duke(-2.5)/GeorgiaTech
15) Stanford(+18)/FloridaSt
16) UCLA(-2.5)/Maryland

The algorithm’s top two selections have gone 9-5 (.643).

This week’s edition of (maybe boring) tech talk is -- weekly line movement. As in, does directional-movement over the five or six preceding days forecast a winner? The logic there is – the ‘collective mood’ of players moving the line ($) has value. Many years-ago I encountered at least two published studies that concluded there was no-edge in doing that watch. However, there’s never any harm in a current data dive to see if anything recent might be lurking.

And of course, a quick peek at the database tells all (2013-24, 7583 games). In doing so, I disregarded the magnitude of the movement; early line versus final line. First off, movement was evenly divided between the Home and Away teams, with only 15% (1134) of games having no movement. When the Away team gained points, it had a Success Rate of .516. Conversely, when it lost points, it’s SR was .510. Thus, it must be reported.......no (lurking) movement-edge to be had.

Sliding laterally on that study though, you might be curious just how often the Away team was victorious in that 11-year base. That would be .511 (3816-3648-119), suggesting a tiny-bit more value in the road team.

We have 59 on the board; once again surpassing the weekly-opportunity of all 2025 weeks past.
Good luck,
TheKingfish

SamHouston(+3)/UTEP

Not at the top, but the algorithm has the Bearkats in the upper echelon this week. Notable, these two programs have combined for a real dearth of cashed tickets in 2025.....

Good luck,
TheKingfish

Adding yet another (the KSU Wildcats), and doing a coherent re-order of the list.
1) Colorado(+4.5)/IowaState
2) Wisconsin(+3.5)/Iowa
3) Arizona(+2)/BrighamYoung
4) KansasSt(+2)/TCU
5) SamHouston(+8.5)/JacksonvilleSt
6) TexasA&M(-7)/Florida
7) LSU(-8)/SouthCarolina
8) Missouri(+3.5)/Alabama
9) MichiganState(-7.5)/UCLA
10) Oregon(-7.5)/Indiana
11) USC(-2.5)/Michigan
12) OregonState(+2.5)/WakeForest
13) Washington(-10.5)/Rutgers

The algorithm also quietly likes both EasternMichigan & GeorgiaSt (in the 5.5 position), but please pencil them in for me only if each moves up to +3. Both are currently hanging at 2 to 2.5. Thank you.

Good luck,
TheKingfish

For the record – am adding another.
2.5) Arizona(+2)/BrighamYoung

Truth be told, I actually have three (3) more waiting in the wings. In common.......the box insists they all need a bit-more incentive (points).
Good luck,
TheKingfish

A mild winner here last Saturday (7-6, .538). Despite just two losing weeks out of six, the full-season record is nothing to write home about (39-37, .513). NG. Turning that crank on the box, and looking at each game, these float to the top:
1) Colorado(+4.5)/IowaState
2) Wisconsin(+3.5)/Iowa
3) SamHouston(+8.5)/JacksonvilleSt
4) TexasA&M(-7)/Florida
5) LSU(-8)/SouthCarolina
6) Missouri(+3.5)/Alabama
7) MichiganState(-7.5)/UCLA
8) Oregon(-7.5)/Indiana
9) USC(-2.5)/Michigan
10) OregonState(+2.5)/WakeForest
11) Washington(-10.5)/Rutgers

Through the first six weeks, the algorithm’s top two selections are performing (8-4, .667). Alas, a small sample size and maybe just that blind-pig rummaging about for his acorns.....

Hey, consistent with this slot on the forum, how about some more boring data -- to satisfy that internal-curiosity of knowing how the categories are faring in 2025? Well, over 292 FBS non-neutral site games, Home Favs are 93-89-7 (.511), and Home Dogs are 50-51-2 (.495). So, we’d all be challenged to find any strategic direction in those numbers. As we know, these percentages all go to .500 over time.

Another dollop of boring; though maybe ominous in nature. SpreadMargin (SM) is the amount a team covers by -- or fails to cover. Over the last eleven seasons (7583 games), the SM was 12.23 points. This season, SM is 11.51; a reduction of six percent. Within our current realm of understanding, and grappling-with, Artificial Intelligence, I leave every capper to worry (or not-worry) about that six (6) percent.

We have 56 on the board this week – the most weekly-opportunity thus far.
Good luck,
TheKingfish

Am adding yet another that was forgotten earlier – with a re-order in order.
1) SanDiegoSt(-5.5)/ColoradoSt
2) Maryland(+6.5)/Washington
3) SouthAlabama(+2.5)/Troy
4) Akron(+8.0)/CentralMich
5) UAB(+7.0)/Army
6) Florida(+7.5)/Texas
7) FloridaSt(+4.5)/ MiamiFlorida
8) Temple(+ 6.5)/UTSA
9) CentralFlorida(+4.5)/Kansas
10) NoIllinois(+5.5)/MiamiOhio
11) NewMexico(+2.5)/SanJoseSt
12) WestKentucky(+2.5)/ Delaware
13) Clemson(-14)/NorthCarolina

As scatterbrained as I seem to be this week, I can’t guarantee there won’t be more....
Good luck,
TheKingfish

For the record -- one more to add.
12) Clemson(-14)/NorthCarolina

Also, for the record; am still holding cash on the Friday night Lobos & Hilltoppers. Put me down for +3 on both…….if the number ever gets there. And, thank you to Mr. JJWoods for the support!
Good luck,
TheKingfish

A winner posted here last Saturday (8-7, .533), bringing the full season to an anemic 32-31 (.508) record. It seems I’m providing little value thus far. As ever, evaluating every game on the board (and this week eschewing the chalk), the box placed all of these on the ticker tape:
1) SanDiegoSt(-5.5)/ColoradoSt
2) Akron(+8.0)/CentralMich
3) UAB(+7.0)/Army
4) Florida(+7.5)/Texas
5) FloridaSt(+4.5)/ MiamiFlorida
6) Temple(+ 6.5)/UTSA
7) CentralFlorida(+4.5)/Kansas
8) NoIllinois (+5.5)/MiamiOhio
9) SouthAlabama(+2.5)/Troy
10) WestKentucky(+2.5)/ Delaware
11) NewMexico(+2.5)/SanJoseSt

Through five weeks, the top two selections have performed profitably (6-4, .600). The tweaked latter-portion of my 2024 season was also profitable -- possibly a reflection of the more-mature power ratings. That was hard to know though, given the small sample size. Nonetheless, the Human Administrator made an algorithm ‘adjustment’ this week; trying to coax some similar sunshine. Nine weeks left in the trenches.....

Hey, how about some boring tech-stuff serving as a shameless distraction? Middles. I attempted one last November, but hadn’t actually tried one since 2021. I have placed seven of them since 2018; hitting just once. That one had some exceptional payoff odds (36/1) so, by that math, I can wade-in about 29 more times before I reach the breakeven ($). Conventionally though, the payoff is more like 24/1, so a revised-future would be less lofty than that 29-attempt cushion.

My MIN gap-criteria for a middle is four (4) points. And of course, one needs to also be on the right side of the play; holding a team that’s grown more attractive at the earlier-placed line. Concluding with some geeky data-collection, it naturally rolls to ask how often that >=4 point movement happens (???). My database captures only the early & final lines, so I can’t account for what happens in between; likely presenting slightly more opportunity.

Anyway, here’s the cheap-thrill data payoff. During seasons 2023-4 (1469 games), there were 71 games with movement of four points or greater (4.8%). This year (242 games) has seen 11 such games (4.5%). That movement is proportionally weighted toward the away team gaining points; roughly twice as often. Bottom line – middles are tough to pull off. However, it’s a totally sweet-gambit for players who are both watchful & nimble. And, did I mention..…..lucky?

We have 50 games on the board this week -- an abundance of opportunity. Enjoy the ride.
Good luck,
TheKingfish

For the record – adding yet another.
17) EasternMichigan(+3.5)/CentralMichigan

Current line shown above. I have not yet placed on this one ($) -- as I’ve watched the Eagles tumble from a high of at least +6.5. Over the years, absent any hard data collected, I’ve seen the number routinely revert to its beginnings. So, I’m watching & waiting; even though 3.5 is an absolutely-stellar Dog number.
Good luck,
TheKingfish

A non-winner posted here last week (6-7, .462). That puts the season overall at .500 (24-24). NG. Tech cappers often have the capacity to examine every game. In doing that, the box favors all of these:
1) SanJoseSt(+2.5)/Stanford
2) AirForce(-4.5)/Hawaii
3) ColoradoSt(-4.5)/WashingtonSt
4) GeorgiaTech(-14.5)/WakeForest
5) LA-Monroe(-1.5)/ArkansasSt
6) FloridaSt(-6.5)/Virginia
7) Northwestern(-6.5)/UCLA
8) Illinois(+6.5)/USC
9) NoIllinois (+3.5)/SanDiegoSt
10) Arkansas(+6.5)/NotreDame
11) Iowa(+7.5)/Indiana
12) SoMississippi(-3.5)/JacksonvilleSt
13) TexasA&M(-6.5)/Auburn
14) Buffalo(+3.5)/Connecticut
15) Army(+4)/EastCarolina
16) Rutgers(+5.5)/Minnesota

Like the Saturday before, the system brought forth a more spaced-out level of success. However, the algorithm continues to show some rank-value, whereby the top two have gone .625 (5-3).

Anecdotally, it feels like the large favs are winning in 2025. At least when recording scores, it has seemed so. Is that true? Easy enough to check. The reality is, my “feels like” is not true. Away Favs giving 11+ have fared poorly (8-11, .421). And Home Favs giving 15+ have similarly not-cashed their slips (27-38, .415).

However, if your gaze shifts to the long-term (many seasons) – it all goes to .500 of course. Which brings us back to the salient point we all know – it’s a tough game to beat. I rotely embrace the favs when so directed (box), but personally more-like the value of dogs. Possessing no fundamental-type handicapping talent – it’s just something that I observe tangentially in my data.

We have 51 games on the board this week; the most so far. So, abundant opportunity for one and all!
Good luck,
TheKingfish

For the record – adding one more.
3.5) WesternKentucky(-8)/Nevada

The algorithm didn’t like the early range (peaking around -11.5), but always liked the Hilltoppers at the magic number of -8. A level I never thought it would reach. In any event, we’ll see how much “magic” that number holds this evening.....
Good luck,
TheKingfish

A winner posted here last week (9-6, .600). That puts the season overall at .514 (18-17). No money to be made on that. In classic tech-fashion (evaluating every game), the box likes this bunch for week four:
1) FloridaInter(-4.5)/Delaware
2) UTEP(-4.5)/LA-Monroe
3) SoMIssissippi(+ 3.5)/LouisTech
4) BoiseState(-9.5)/AirForce
5) EasternMich(+3.5)/LA-Lafayette
6) Hawaii(+3)/FresnoState
7) Buffalo(-6.5)/Troy
8) EastCarolina(+7.5)/BrighamYoung
9) Charlotte(+3.5)/Rice
10) MiamiFlorida(-7.5)/Florida
11) Arkansas(-6.5)/Memphis
12) NorthTexas(+1.5)/Army

Last Saturday the system brought a broader level of success, but failed to perform at the top. Nonetheless, it has so-far shown rank-value for the top two picks (5-1, .833).

The three-week mark is pivotal, in that the Human Administrator feels there’s enough data to assess the system internally for its cutoff points. It’s a small sample size, but that dutiful weekly-exam will continue on into the season.

By design, the algorithm is structured to produce volume; while realizing a reasonable success rate. That combo has intrinsic value – which may, or may-not, be achieved. During the final eight weeks of 2024, the top eleven went .650 (52-28-3). In other words, the model has some positive history. It’s still early, so only the wait & see will show how much “value” the 2025 algorithm can generate.

We have 50 games on the board this week; ample opportunity for all.
Good luck,
TheKingfish

For the record, based on a moving line -- I need to sneak in one more.
2.5) JacksonvilleSt(+3)/GeorgiaSouthern

Good luck,
TheKingfish

A winner was posted here last week (6-4, .600). That takes the season overall to .450 (9-11). Not so good. However, looking at every game, the box likes these for week number three:
1) FlaAtl(+2.5)/FlaInter
2) NorthTexas(-5.5)/WashingtonSt
3) Liberty(-6.5)/BowlingGreen
4) CoastalCar(+6.5)/EastCarolina
5) NorthCarSt(-7)/WakeForest
6) WestVirginia(+7)/Pittsburgh
7) UtahState (4)/AirForce
8) GeorgiaTech(+4 )/Clemson
9) Duke(+3)/Tulane
10) Vanderbilt(+5.5)/SouthCarolina
11) Kentucky(-24.5)/EasternMich
12) LSU(-7.5)/Florida
13) MiamiFlorida(-16)/SouthFlorida
14) UCLA(-15)/NewMexico

So far, the system has demonstrated some rank value (top three picks, 5-1, .833). That certainly won’t last, but it’s worth basking-in for the moment.

Being a Power Rating player, I did some research in the off-season regarding the Delta; the weekly amount of change. There are a number of PR sources (Steele, Sagarin, Massey, VSIN, SP+, more), so you can pick your poison. I’m guessing the source would not appreciably alter the research outcome -- but difficult to research them all.

Anyway, change slowed going into week #4; stabilizing for the remainder of the season. That makes sense, and reflects a seasonal database that grows to the point of a diminishing return. That would suggest a degree of predictive volatility in the early going. A few more courses would have garnered a math major – so I’ll defer any more comment until I get around to completing those.

We have 47 on the board this week, so plenty of opportunity!
Good luck,
TheKingfish

@River_fish

For the record. Based on a moving (and slightly more-attractive) line, let me sneak in one more.
5.5) Marshall(-7)/MissouriSt

The algorithm held some affection for the Herd, but not at the earlier -10 and so forth.
Good luck,
TheKingfish

Some poor performance here last week (3-6, .333). Which brings the young season overall to .300 (3-7). Not good. Dutifully feeding in data for a look at every game, the box likes these:
1) NorthCarolina(-13.5)/Charlotte
2) Hawaii(-7)/SamHouston
3) TexasSt(+4.5)/UTSA
4) FresnoSt(+3)/OrergonSt
5) Virginia(+3)/NorthCarolinaSt
6) Baylor(+3)/SMU
7) ArizonaSt(-6.5)/MississippiSt
8) Colorado(-23.5)/Delaware
9) Florida(-17.5)/SouthFlorida

The algorithm also likes a handful of dogs not listed above (Duke/SanDiegoSt/UNLV). However, they all need to get-to a certain upward line (3.5/2.5/3.5). A few years ago, I incorporated an internal mechanism with the scope to assess lines higher & lower than present magnitude – allowing me to loosely anticipate. Very helpful to the process....

Apart from having a disappointing overall week, last Saturday did hold some rank-value (top five 3-2, .600). My system routinely delivers volume, so rank value can be a by-product of that. In fact, I have never really had a handicapping year where progressive rank was not imparted. Now, precisely where that seasonal magic-cutoff falls -- remains a mystery until December. Top four? Top six? Eight?

Even the Human Administrator is blind to that aspect! The current good news is -- we have 50 on the board this week; lots of opportunity.
Good luck,
TheKingfish

An unspectacular small-start to the season here last week (0-1). Giving the crank a hard full-turn, the box generates these:
1) California(+2)/OregonSt
2) Tulane(-4.5)/Northwestern
3) UCLA(+6.5)/Utah
4) Texas(+2.5)/OhioSt
5) GeorgiaTech(-4.5)/Colorado
6) SanJoseSt (-11.5)/CentralMichigan
7) GeorgiaSouthern(+1)/FresnoSt
8) CentralFlorida(-18.5)/JacksonvilleSt
9) MichiganSt(-20)/WesternMichigan

The algorithm also likes Baylor, but not at the present +2.5. Having a rather truculent mind of its own, it wants the Bears at >= 3....

Given my tech MO, I evaluate every game on the board – with the hope that my shared-season performance will prove positive. Alas, early optimism is cheap! Good or bad outcome though; I’m here for the duration.

We have 46 on the board this week, so an abundance of opportunity for all.
Good luck,
TheKingfish

Glad to return for a fourth year on the forum! With hope springing eternal, my resume is unchanged. I’m low on flamboyance, with a capping style that is decidedly technical. I am a Pascal programmer; a number-cruncher and a slave to data/history. I use a self-created algorithm comprised of historic trends culled from a 2013-24 database (7,583 games). Foundationally, my algorithm uses Power-Ratings.

Years ago, I used a generic algorithm (no PRs). It was less successful for me, but held a powerful amount of data (1976-2012 / 19,588 games). During this past off-season I did some research and added an adjunct wing to my executable. I’d seen evidence that time zone travel affects ATS performance in all sports, so my curiosity was peaked for the NCAAF side of it.

Given that long-travel (coast to coast) is far less frequent than otherwise, I felt that only my old very-large database might be effective in analyzing that smaller sample size. My need, of course, was to quantify. So, does such phenomena exist in college ball? My research says it does – more pronounced for teams going west-to-east than the reverse. And the output data consistently favors the home team; as any good researcher would hope to discover. Or, better said: Hypothesis always precedes Experiment within the realm of good scientific method.

Anyway. In the big picture, if history repeats, there is hope for my system. OTOH, if the future diverges from its noble-past -- then success becomes more elusive for me.

Full disclosure on 2024 performance – my first-half was less than profitable. However, I tweaked the system (adjusted the algorithm) and brought some value during the final eight weeks (top eleven picks, .650, 52-28-3). Gratifying to impart that.....

However, that was then. This week we have 4 games on the board; next week 46 games. So, in technical terms, we will soon be awash in opportunity. For the record, I am sharing my first Saturday selection: Stanford(+2.5)/Hawaii

As a tech guy, I tend to bring volume each week. And volume is a positive when linked with reasonable success. Albeit, my overall geek approach will have limited-interest for many. In the end, I hope to make a positive contribution to the forum, and wish all of you well in the challenging new season ahead.
Good luck,
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