From

Location

Message

Date Posted

Deer Park, WI

You said what?!

I know, I know, this is like poking the bear and swatting the bee hive all rolled into one big ball of fun. I really enjoy stirring the pot.

Mr. McMullen's creation is absolutely the hottest commodity in our hobby. And I'm not taking away from the fact that it has produced the most notable pumpkins of all time. It has the best track record of production.

I take issue, however, with the label that this seed is the "best" seed out there, when looked at through the lens of genetic potential.

Those who have read my commentary on seed genetics in the past will recall that I favor a rudimentary, back of the envelope system of semi-quantitative analysis called "cross weight average (CWA)"

You can find a thread on this topic from a year and a half ago here, with its definitions and well stated constructive crtiticisms: http://www.bigpumpkins.com/MsgBoard/ViewThread.asp?b=3&p=561371

I have compiled CWA analysis on both 2016 and 2017 GPC top 20 pumpkins in a short article just posted on the SCGA website. All genetics enthusiasts can read this article for free on the homepage: www.stcroixgrowers.org

In a couple weeks, it will be moved inside the members-only content section of the site.

The 2145 McMullen CWA is (2145+1756+1625)/3=1842. There is quite a spread between this value and the #1 CWA seed, 2150 Skinner (2624 x 2624) = 2466 CWA.

I will not be so bold as to say that the 2150 Skinner is a better seed than the 2145. We simply do not know. However, if I was asked to choose which seed to grow, in the absence of historical performance data, the CWA formula would suggest that the 2150 seed may have greater potential.

11/23/2017 9:08:29 AM

Deer Park, WI

Critics may rightly say "but the 2145 has a proven record". Sure, it certainly does. But its not fair to compare potential to 2150, which presently does not have a record. And very well could have a record that supersedes 2145.

Here's my take away- ask 10 growers on this site what seed would be their first choice to plant next year and 11 or 12 of them would say 2145. Not everyone is going to get one tho. In lieu of that, use what tools you have to help make decisions that position you to optimize genetic potential. It is my belief that CWA is the most effective way to do so.

Critics, fire away!

11/23/2017 9:08:38 AM

Muddy Waters

I respect what you are saying Joe & totally understand it.But with low seed count on Todds I dont think we will find out. Only way is if you get 10 Heavy hitters to grow it.Grower error plays a big role in proving a seed.Many new growers will get a hot seed & not realize the importance of the soil being spot on & weather & so many variables.A lot of silver bullets are lost to grower error.


What is the value of my 2043.5 =2145X 1947.5/3 2045. Is this correct? Even though it only went chart (or 17lbs under)

11/23/2017 9:40:29 AM

Muddy Waters

sorry (2043.5+2145+1947.5)/3 = 2045 lbs ???correct Joe?

11/23/2017 9:43:07 AM

Muddy Waters

So my 2043 should be superior the the 2145??? numbers wise? or am I missing the bus on this?

11/23/2017 9:44:14 AM

Eddyz@efirehose.net

Joe
I've kinda been doing this number routine also, but I think a handicap should be added for light to chart vs Heavy
Also does the fact of the self in the Skinner pumpkin qualify for 2 2624's or should it actually only be 1 2624.

Other seeds of note with this formula
2269 Patton 2269+2624+1949 Average 2280
2363 Holland 2363+2145.5+2261.5 Average 2256.66
2624 Mathias 2624+2145+1872 Average 2213.8

11/23/2017 10:10:56 AM

Massachusetts

If my notes from Niagara are correct the 3 pumpkins that the system identified as the top 3 seeds to grow this season were the 1944 Connolly, 1937 Urena and 1938 Zywiec. Looking at the top 100 GPC the system had 5-6 pumpkins in the top 100 pumpkins grown. Is this any better really than what an average grower can pick by looking/talking/guessing?

(My apologies if my Niagara notes are not accurate. I am reading from a napkin with pencil scratching.)

11/23/2017 10:37:23 AM

Netherlands

Tom, it all depends on how often the CWA top 3 were planted in comparison to the 2145 and by which growers.
The CWA considers possible potential and is no reflection of what happens in real life. It does not account for the number of seeds distributed or actually planted or possible "mutations" in one of the parents.
Eddy makes a good point about selfed pumpkins. We also all know that WR pumpkins seldom produce new WRs. CWA is simply a tool (along with other tools) to help choose a seed without a track record.
Mathias Willemijns proved last year and this year that you can grow 2000lb pumpkins from pretty much any seed which means that the possible "best" seed out there has never been grown.

11/23/2017 10:57:05 AM

Deer Park, WI

Bubba, your math is correct. Here's the link to the article where you can see how it stacks up relative to the other 19 pumpkins in this year's GPC top 20: http://www.stcroixgrowers.org/cross_weight_average.pdf

I will not say your 2043 is superior to 2145 because it has a higher CWA, rather, in the absence of any historical information and given a choice between the two, the equation would favor your seed.

Eddy- Todd pollinated his 2624 plant with 2624 pollen. as such, he has created a cross that is the heaviest grown by heaviest grown. I have chosen not to adjust the formula based on self/sib crosses. Todd's 2150 CWA therefore is (2150+2624+2624)/3=2466. There's no reason to believe his cross is at any genetic disadvantage because he selfed it. In fact, quite the opposite in my opinion.

As far as a handicap for % light/heavy, I too wish we had a way to quantitatively favor genetic lines where known wall density/thickness is above average. It would be incredibly difficult to derive a handicap value that could be useful, due to variance and all the factors that contribute to this characteristic. Simply applying an arbitrary value based on how the actual fruit weighed up or down on the scale seems out of reach.

Handicaps also open pandoras box. Growers +/- X number of degrees off the 45th parallel should be handicap for less than ideal growing seasons, growers with greenhouses...hailstorms/torandos/hurricanes....and so on.

11/23/2017 11:10:19 AM

Deer Park, WI

Beautifully stated, Brad!

11/23/2017 11:11:39 AM

Muddy Waters

Thanks Joe I learn a lot from you.

11/23/2017 12:31:00 PM

Santa Rosa, CA.

If you could handicap my 12% heavy 1938 grown from a 2230 Wallace and pollinated with the 2066 Geddes that also went 12% heavy. Boom!!!

11/23/2017 12:58:33 PM

Torrance, Ca.

I have a small handful of those 1938's in the vault!:) That's going to be the next big thing, mark my words!

11/23/2017 2:13:41 PM

Eddyz@efirehose.net

If another pumpkin seed was grown as extensively as the 2145, would the results be the same, I agree that it would, could be better, could be worse, they just need to be planted.

11/23/2017 3:19:04 PM

Colorado

Another way to rate a seed is did it grow a PB and by how much.The 2145 helped a lot of growers get into the 2000 lb club /grow them a Personnal best and it a few cases they smashed their PB's.Saying its not the best seed is just silly.

11/23/2017 5:28:04 PM

NE Wisconsin

Pretty good point Cojoe. Would factor in growers ability/location. Not every grower has tried all of the top seeds probably. Start a "what seed grew your personal best thread. If people list their top 5/10 you could sort of compare. Still a million other factors that could affect each different plant though.

11/23/2017 6:02:58 PM

nOVA sCOTIA

I know a grower who has never grown anything but proven seed and still can't grow a record pumpkin so explain that one..lol

11/23/2017 7:16:32 PM

Anamosa, IA BPIowegian@aol.com

Big thing to remember is that every seed from a pumpkin will have slightly different genetics. Each one an individual with its own potential. So you could get some real giants and some runts from the same pumpkin. As we keep selectively breeding, hopefully we will reduce the percentage of runts over time. Maybe what we need to do is self pollinate several seeds from the same pumpkin, and grow only seeds from giants. maybe you can eliminate the runt genetics over time.

And you can't discount other factors in how big a pumpkin gets. I may have my fertility at optimum levels, but my cool, humid valley will limit me. I miss the early morning and late afternoon sun, so I don't expect to ever obtain world record size. But I can at least breed pumplins for other factors, such as shape and color.

11/23/2017 7:55:27 PM

Owensboro, Ky.

Kudo's to cojoe!!! Not positive but thinkin that the VanHook's grew a couple of 2145's to new personal best's (in Kentucky) & massive % OTC! type of statistics. Am confident that they will do so again next year, with same/other seed !!! They don't stop doing it better!! Slungers!!! Peace, Wayne

11/23/2017 10:55:04 PM

Wisconsin

I thinking throw a few darts at the last couple years top 100.

11/24/2017 12:55:19 AM

Germany

Let the darts hit the envelope, not the seed ;)

11/24/2017 1:07:43 AM

Buffalo, MN (Billsbigpumpkins@hotmail.com)

Joe, I like your topic and thoughts. The best way to know if the 2145 is a better seed than any other seed is to grow 10-25 of each seed in exactly the same controlled condition. The second best is to use all the seeds the last several year’s statics and get the average weight the seed produced. If there is less than 20 seeds used it noteworthy but not yet a champion seed. Yes you can send me you extra 2145.

11/24/2017 10:08:33 AM

Willamette Valley, Oregon

A few items to address:

sample size or number grown MUST be taken into account when making a fair assessment -- the total number of fruit grown plays a huge roll in the apparent success of a seed. A seed as widely grown as the 2145 (I find 75+ grown in 2017) has the clear probability advantage in terms of the growing the largest fruit AND as cojoe mentions being a lot of peoples' top seed. If you bring either of those back to proportion of total grown it becomes much less clear that the 2145 is the hands down winner.

Analogy: So pretend there are three baseball players.
Jim hit 25 home runs in a year and Joe hit 50 home runs that year. Who's a better home run hitter?

Now suppose I told you Jim had 125 at bats and Joe had 500 at bats. Wanna change your answer? (You should.)

11/24/2017 9:44:20 PM

Willamette Valley, Oregon

Grower ability must also be taken into account, but how to do that mathematically is much less simple -- Many heavy hitters grew multiple 2145s so that certainly boosted it's chances of being heavy hitter's top seed.

Analogy 2: Jim and Joe again, same number of home runs as before: 25 and 50. Let's say this time they had the same number of at bats, let's say 250 each.... now who's a better home run hitter?

Except this time Jim is a major league hitter (with better pitchers aka novice growers) and Joe is a minor league hitter (with worse pitching aka more experienced growers).

Now who's the better home run hitter?

11/24/2017 9:56:22 PM

Willamette Valley, Oregon

Analogy 2 could also apply to weather or other factors that might effect the seeds performance besides the seed fundamental quality.

11/24/2017 9:59:43 PM

Willamette Valley, Oregon

ok let's discuss heavy-to-chart vs light-to-chart seeds:
This one I have analyzed. On a parent seed basis there is actually not a strong relationship between the average %-to-chart and the average offspring weight. But the data hint that the seeds that produce the highest average weights are those that are closer to 0% to chart (+/-). Beyond that, to the + and minus side the average weights decrease. That means that heavy-to-chart seeds may not be your best choice if your ultimate goal is the weightiest pumpkin at the weigh-off.

I'll grant you that it is nice to "go heavy" on weigh off day but ultimately pounds is pounds...and the most pounds wins. I would chose a 20% light 2700 lber over a 20% heavy 2690 lber every time.

11/24/2017 10:11:59 PM

Willamette Valley, Oregon

Getting back to Joze's assessment method:

It's not terrible as a quick and dirty assessment, but... To understand a seeds potential you must know the genetic father's traits (and probably the grandfathers') all 3 of which are ignored in that method. And by genetic father I don't mean the seed that grew the father plant. Let me explain...

consider this standard seed label:
2624.6 Willemijns 2016 (2145.5 McMullen 2015 x 1872.8 Willemijns 2015)

First thing to note is that the 2624.6 seed and the fruit that produced it, the 2624.6 fruit, are called the same thing, but in human terms that would be equivalent to calling me by my mom's name. It's terribly confusing.

So...from the genetic perspective of the 2624.6 SEED, its mother is the 2624 FRUIT. 2145.5 FRUIT is the 2624.6 seed's maternal grandma. The 1872.8 FRUIT is the 2624.6 seed's paternal grandma. That's a pretty limited genetic history to be basing assessments on. It IS what's readily available though, so that's a plus.

11/24/2017 10:27:41 PM

Willamette Valley, Oregon

With a little sleuthing one can determine that the genetic father of the 2624.6 seed is the 1872.8 offspring plant that produced the 2063.5 Willemijns 2016 Fruit.

Knowing the genetic father is key to assessing the quality of a given seed (really, it's vastly more useful than knowing the maternal and paternal grandmas).

So I'm strongly in favor of seed labels that look like this:
2624.6 Willemijns 2016 (2145.5 McMullen 2015 x 1872.8 Willemijns 2015 [2063.5 Willemijns 2016])
where the true genetic father plant is recognized within the brackets as [2063.5 Willemijns 2016].

(Note: genetic grandfathers are still ignored, and these can be extremely important, for example, in squash breeding.)

Armed with the extremely useful value of Father weight, a better assessment of seed quality is the (shazzy coined?) "power on both sides" method. Nothing tricky: Genetic mother + genetic father.
So the 2624.6 value is 2624.6 + 2063.5= 4688.1.

This still faces most the same criticisms that Joze's method does, but I feel it's closer to the true genetic potential.

Some notable seeds this method would have identified earlier are:

1161 Rodonis 2007 (1231.3 Pukos 2005 x 1450 Wallace 2006 [1566 Rodonis 2007])

1385.5 Jutras 2007 (1068 Wallace 2003 x 998.6 Pukos 2005 [1689 Jutras 2007])

But also many other recent seeds that were vastly under planted (in my opinion).

11/24/2017 10:50:37 PM

Willamette Valley, Oregon

And lastly - a note on Selfing and selfed seeds.
In short, selfing is awesome. It ensures that the best fruit are crossed with the best fruit. And helps to lock in the best genes going forward rather than diluting them like out crossing would do.

11/24/2017 10:54:19 PM

Willamette Valley, Oregon

Gotta love those selfed Skinner seeds and of course the 2002 Tobeck 2017 (2145 x self).

11/24/2017 11:44:31 PM

Grittyville, WA

Good job stirring the pot Joe!

11/25/2017 2:47:48 AM

Grittyville, WA

@ cojoe and Wildcat 83... I'm working on it, it's coming.

Its taking a long time because its not the easiest method (sorting data by PB's...). The Mc M. looks good of course, its a very good seed, but hopefully at the end I'll have a list of seeds that might be better.

Good posts bathabitat... I can tell you are smarter than me.

11/25/2017 7:12:05 AM

Grittyville, WA

Lol of course that doesn't say much because i'd say about 2/3 of the people on here are smarter than me...

Now, based on that limited data, what is the exact % chance that bathabitat is smarter than the top 50% of everyone on here? First person to answer gets... pgvg club seeds/field kins/exotic seeds/2015 holland dvd (I'm scraping the bottom of my refrigerator for you guys, maybe i'll even wash the gummy worms and potato chips off the dvd)

11/25/2017 7:53:59 AM

Deer Park, WI

Mr. Bathabitat, your posts are wonderfully articulated. I have also long believed that seed nomenclature needs an upgrade (exactly as you described) and that a 4-way cross weight average would be a more powerful indicator to semi-quantitatively assess seed genetics.

11/25/2017 8:11:01 AM

Owensboro, Ky.

Mr. bat sir...and no disrespect meant!!! Can you do a research thing on the % over chart on VanHook pkns, for the last 3 or 4 years? I think all would be overwhelmed by the results!! Granted, they are not record setters...as they grow in the mid-south! But results...% wise are very interesting! Peace, Wayne
Curious...breeding has to be important...as opposed to growing conditions!

11/26/2017 12:35:21 AM

Grittyville, WA

He's a top source of real info.

11/27/2017 3:18:27 AM

Willamette Valley, Oregon

VanHook offspring do seem to be consistently heavy. Four fruit averaging 15.25% heavy in 2017.

Picking heavy genetics might get a grower something like +3 to +7% on average, so consistently higher than that is either very lucky or he's doing something to affect it.

For example the 2145.5 McMullen offspring averaged about 2.7% heavy in 2017.

11/27/2017 4:24:29 PM

Muddy Waters

Iowaegian is spot on every proven seed has runts .This is disappointing for many. My 1756 this year didn’t produce the pumpkin started got bowling ball size and pretty much stopped it didn’t abort it just grew so slowly might as well of stopped So there is the luck of the seed factor.But I think a lot of people are missing the fact that every seed in a pumpkin has a different chromosome line up just like siblings in a family or snowflakes no 2 are alike

11/29/2017 10:28:30 PM

Willamette Valley, Oregon

cojoe - you're gonna love this one:

So I calculated the number of times a seed was the best in an individual grower's patch (for growers who got two or more fruit to the scale) relative to the number of times it the seed made it to the scale anywhere, and...

of all the seeds grown 4 or more times total...

the best seed by that metric at a rate of 75%...:

1685 Scherber, 3 were grower's best in the patch, out of 4 fruit total that hit the scale.

800 est McMullen - 10th place with 50% (2/4)
2145.5 McMullen 15th place, with 41% (31/76)
2230.5 Wallace 16th place-tied with 40% (4/10)

*disclamer - I wouldn't recommend using this metric to pick seeds, but it's still fun.

11/30/2017 6:30:56 PM

Willamette Valley, Oregon

Rate of 2k+ production - That is, the number of times a seed grew a 2000 lber divided by the total number making the scale (for seeds hitting scales 4x or more):

Rank, 2kRate, Total_weighed, SeedName
1 0.25 04 1949 Patton
2 0.20 05 2009 Wallace
3 0.17 06 1810 Werner
4 0.15 13 2624.6 Willemijns
5 0.11 76 2145.5 McMullen
6 0.10 10 1937 Urena
7 0.04 23 1844.5* Holub
8 0.04 24 2095.5 Willemijns

-disclaimer as above.

11/30/2017 6:41:41 PM

Deer Park, WI

Fascinating!!!

12/1/2017 6:47:37 PM

Little-TON - Colorado

Agreed....

12/2/2017 11:01:06 AM

Willamette Valley, Oregon

Here's that 2k+ Rate for all the seeds that grew one over 2k and grew 4 or more total ever. (my post above was just 2017 fruit)

2kRate TotalGrown Seed
0.25 004 1872.8 Willemijns 2015
0.25 004 1949 Patton 2016
0.20 005 1861.8 Cutrupi 2015
0.17 006 1385.5 Howell/Jolivette 2014
0.17 006 1810 Werner 2016
0.15 013 2624.6 Willemijns 2016
0.13 008 1961.5 Sperry 2015
0.12 017 2096.6 Meier 2014
0.11 018 1585.5 Werner 2014
0.11 009 1894 Mathison 2013
0.10 108 2145.5 McMullen 2015 ~~~~~~~
0.10 010 1937 Urena 2016
0.09 022 1338 Martin 2013
0.06 125 2009 Wallace 2012
0.06 016 1781 Zywiec 2014
0.04 023 1844.5* Holub 2016
0.04 023 2036 Glacier 2014
0.04 024 1975 Wallace 2015
0.04 024 2095.5 Willemijns 2016
0.03 030 1756 Howell/Jolivette 2014
0.02 041 1730.5 Werner 2012
0.02 100 1985 Miller 2013
0.02 053 2032 Mathison 2013
0.01 075 1725 Harp 2009

A rate of 0.05 (1 in 20) to 0.15 (3 in 20) seems to be typical of "good" semi-recent seeds.

High rates with 4-5-6 fruit are probably by chance. Same with some of the lower rates with few fruit. This list could also be much longer but many fruit are probably excluded by bad luck or not enough offspring.

The point is that 2145, while probably "good", is not extraordinary, other than it was planted a lot.

12/4/2017 11:31:48 PM

Willamette Valley, Oregon

For comparison...

The 2k+ rate in the overall 2017 population of fruit is around 0.01.

Looking at a set of top growers, in this case the top 40 GOTY candidates...that is those growers that got 3 or more fruit to the scale...when looking at all the fruit they grew and weighed, not just the top 3 (156 total), the 2k+ rate was about 0.06.

I'm not sure what that tells us actually, but it seemed like an interesting comparison.

12/5/2017 1:11:45 AM

Streator ,Illinois

Baseball quite the analogy:

Jim had 25 homeruns and lets say 100 strike outs in 125 at bats and Joe had 50 home runs and 200 singles ,100 doubles, 100 triples and 1 grand slam, 49 strike outs in 500 at bats now choose .

The 1st year out the 2145 hit the grand slam and a home run and some triples and doubles some singles and strike outs as well as being in the minor league, 2nd year now all the major league teams want him and his homerun production triples, with the other hits and strike outs , 3rd year is coming and production is to be seen win or lose the 2145 in some eyes is probably just a good seed has become a game changer for personal bests for others
To try and compare seeds from different years is a shot in the dark , so many seeds get planted that do not make it to the scale , no real scientific data to do a comparison ,
I think of us average growers choose a seed that they think can get them to the next level and are easy to get and have a lot of confidence in seed that they choose. Genetics in any seed you choose is crap shoot, so to use any method to pick a seed that has scientific quantitative analogy is in my mind BS . way to many variables, oh and to say that the 2145 is probably just a good seed is full S*^t its much better then good
Gene

12/5/2017 5:33:50 AM

Somerset,Ky

I'm with you Gene! Thanks for a super seed. It broke the KY state record twice this year... Thanks for sharing such a super seed!

12/5/2017 8:13:18 AM

Grittyville, WA

I get Gene's point which is the 2145 is not a "middle of the lineup" seed. People do need to be wary of stats because the quantity and the quality of the data going in to the stats is so low. Willemjins hit a grand slam and Holland did too. I would be a little surprised if bathabitat would really argue that its a middle of the road seed... but maybe his analysis method is valid for him.

Gene my own "most doubting" assessment would put it in the top 5 currently available seeds. But I know even a very goood assesment can sound like blasphemy and I realize its probably be terribly annoying. This is how it goes in pumpkinland... We sit around all day second guessing things.

Gene's a bigger grower & better grower & smarter and more generous than the average pumpkin monkey. He offered me a 1756 H/J out of nowhere. He's a big grower but super kind & humble.

12/5/2017 3:49:06 PM

Willamette Valley, Oregon

Gene, Certainly not try to pick on you personally. Sorry if I slighted you. I should have said ""good", by that particular metric." That wasn't meant to be an overarching assessment of the 2145 or any other seed. Any seed that makes it on that 2k+ list, by definition can produce 2k+ fruit, and that’s better than “good” for sure. (Understanding that many other seeds might also be on that list if they were planted more.)

I also should have put my disclaimer on that last post too: This is not the sole metric I'd recommend using to pick the top seed(s). This is just one way to look at the numbers. None of the calculations I've shown is the definitive answer, or maybe not even a good answer. (But they are decidedly better, math-wise, than metrics that don’t take the total number grown into account.)

I agree with you that we really can't tell which seeds are the best with the limited data we have. But I’ll politely point out that everyone uses those numbers. The numbers are really all that we have to go on. And for folks that want to use the numbers, it’s important to be using appropriate math.

12/5/2017 3:57:33 PM

Willamette Valley, Oregon

If good math and sound statistical assessments aren’t adequate to give us some answers, than what does that say about bad math and unfair assessments? I just want to encourage people to use the numbers in a fair and unbiased way and take the answers we get with a big grain of salt that is the recognition that the data themselves are probably flawed in many ways. I don’t think anyone would say that bad assessments of questionable data will somehow be better than good assessments of questionable data.

Interestingly, the 2k+ rates for the 2145 in 2016 and 2017 were both about 0.1. The total production increased because the number grown increased. That is entirely the point of the 2k+ rate calculations. I stand by that assessment, for what it’s worth (which might not be much). Any metric that is more or less constant from year to year is probably a fair one. In any fair assessment the general popularity of the seed, ie, the number grown, must be accounted for to be useful.

I do like Gene’s baseball analogy, because it makes a great point. Even if we knew the data were good, the answer may depend on the metric. If we need homeruns, Jim is clearly the player to put in the game! (at any given at-bat he hits home runs 20% of the time vs Joe’s 10.2%). BUT, maybe homerun percentage (or the 2k+ rate) isn't the right metric for everyone. Maybe it's batting average, say; they just want to get on base. The choice of metric would depend on the situation and goal.

12/5/2017 3:58:43 PM

Grittyville, WA

My apologies for any offence to Gene, also for any offence to anyone else --brandon

12/5/2017 4:14:52 PM

Grittyville, WA

@ bathabitat... the situation is... we have to guess what seed has potential--potential that we've either not yet seen, or potential that we can build upon? and the goal is to reach 3,000 lbs?? great ideas about how to think about things

12/5/2017 4:31:08 PM

stoughton WI

Not to offend any one but there seems to be a flaw in the math, if I'm reading it correctly. It says statistically the 1385 h/j is better than the 1756 h/j. While the 1385 is a really good seed, I know for a certainty that the 1756 is a much better seed and more likely to grow a 2k pumpkin. And in my opinion the 2145 will be difficult to beat until the next great seed is found.

12/5/2017 5:25:18 PM

Gays Mills, Wisconsin

it can be argued that: by statistical data, no seed can outgrow the 2145 :)

12/5/2017 5:50:26 PM

Willamette Valley, Oregon

Rick, I'm not actually trying to say the 1385 is better than the 2145, it's more like a tie on the 2k+ rate metric.

What's your metric cornhusk? ;)

12/5/2017 6:12:03 PM

Willamette Valley, Oregon

Congrats to John VanHook and other state winners!

Not to beat a dead horse, but since the implication is that the 2145 is great because it has won many states this year. I looked into it. One might expect that being grown a lot would increase the chances of a state win and that's true. Taking that into account here's the ratio of state/province wins to the total number grown in 2017:

StateWins StateRatio Seed(Mother)
03 0.43 1756 Howell/Jolivette
02 0.33 1338 Martin
03 0.30 2230.5 Wallace
04 0.29 2066 Geddes
01 0.25 1642.5 Wallace
01 0.25 1685 Scherber
01 0.25 1781 Zywiec
01 0.25 1949 Patton
01 0.25 2063 Willemijns
01 0.25 2106 Schmit
03 0.23 2624.6 Willemijns
01 0.20 2009 Wallace
15 0.20 2145.5 McMullen ~~~~~~~~~
02 0.17 1585.5 Werner
01 0.17 1553.5 Weston
01 0.17 1916 Barron
02 0.15 1800.3 Hoelke
03 0.14 2261.5 Wallace
01 0.14 1795.5 Mccracken
03 0.13 2095.5 Willemijns
01 0.11 1790.5 Wallace
01 0.11 2004 VanderWielen
01 0.10 1801.2 Berrens
01 0.10 1937 Urena
01 0.08 1682.1 Lueders
01 0.04 1844.5* Holub

- standard disclaimer... it’s just one metric, low sample size, etc.

Among top seeds, the expected ratio of state wins to total grown appears to be between 0.1 and 0.3.

12/5/2017 7:15:05 PM

Gays Mills, Wisconsin

of the 1000's of seeds of various genetics only the 2145 has grown the two heaviest fruit of all time! 2145: score two, all others: score zero
since all others have been planted many times more than the 2145 ever has it proves by this metric that it can't be outgrown! .......these types of statements based upon the GPC result data being used is very weak because one can spin/manipulate assumptions on data that holds no strength given all the variables involved. Gene's " to pick a seed that has scientific quantitative analogy is in my mind BS . way to many variables" wraps up this thread nicely

12/5/2017 8:20:48 PM

Grittyville, WA

Yeah ditto on that Cornhusk... Gene's 2145 is the best seed out there as far as we know...

...but if anyone can find a better seed and show that its better in the field my bet is on bathabitat... He's done his homework already. I'm impressed with amount of stat work here and I know this is coming from the guy whose squash went to 2118 lbs. Maybe he is good at what he does and knows what he is doing... My hat is off to all...

12/5/2017 10:05:28 PM

Grittyville, WA

Those seeds look like good seeds worth paying attention to, and worth supporting a club auction to obtain. Good post bathabitat.

12/5/2017 10:15:15 PM

Little-TON - Colorado

Once Gone, New Come, ONE Will..... there is AlwayZ another

12/5/2017 11:02:11 PM

Willamette Valley, Oregon

Cornhusk. I have to point out that your final argument invalidates your metric inasmuch as it invalidates mine. If spinning and manipulating data is the problem, why would your spin and manipulation be a acceptable and mine not? Bs is bs is it not?

Of course I have to conceded it's true that 2% of 2145 seeds have produced two winners. I'm still not compelled to think all 2145 seeds as a whole are hand down the best, when so many other metrics indicate otherwise.

12/5/2017 11:32:53 PM

Owensboro, Ky.

So, WiZZy, what is your guess as to the next "another"? Peace, Wayne

12/6/2017 12:33:19 AM

Germany

One needs internal referencing. One cannot compare the quality of the 2145 seed grown by Matthias Willemijns with a different seed grown by somebody else under different conditions. For internal referencing the "best of patch" approach is a good start (in my opinion), which would then also make other seeds´ quality look representative even if they haven´t been grown by HHs. For example, the 1223 Radach doesn´t look like a surprisingly great pumpkin (when compared with other pumpkins grown off the 2009 Wallace or other pumpkings grown by HHs), but internal referencing (same grower, same year) clearly shows that there´s something special about that fruit grown from the 2009 seed. Hence, one should perhaps eliminate some outliers (damaged plants, some handicaps relative to the other plants in the patch which cannot be attributed to genetics) and then, for the remaining pumpkins, one should assign a score in terms of % above/below average (average = average weight from that particular patch/grower in that year). In this case, even pumpkins grown by backyard gardeners in their rookie year would make useful contributions to the statistics. Just imagine a new grower has limited space in his/her backyard but decides to grow 4 plants...maybe 200 sqft each, and the results are 750 lbs, 550 lbs, 500 lbs, 430 lbs (no outliers, all healthy plants). Average weight is 557.5 lbs, and the scores of the pumpkins would be +34.5%, -1.4%, -11.3%, -22.9%. In case of Matthias Willemijns´ pumpkins from 2016 (2624.6 lbs, 2095.5 lbs, 2063.5 lbs, 2048.1 lbs; average: 2207.9 lbs) the scores would be +18.9%, -5.1%, -6.5%, -7.2%.

12/6/2017 3:56:53 AM

Germany

The biggest problem with this approach is: One needs representative data (the growers who contribute to these statistics would have to deliver unbiased data, weights of all "good" pumpkins grown in their patch in one year, not just picking a selection of the heaviest pumpkins which finally went to the weigh-off, because the smaller pumpkins actually contribute to that measure).

12/6/2017 3:56:59 AM

Germany

Of course, the data would still be biased. Just imagine some growers grow for orange and good shape only, hence the won´t have any of the big and ugly ones in their sets for intzernal referencing (they sort of systematically eliminate the big and ugly ones by deliberately not-growing them). This would create a sub-group of internally referenced pumpkin seed productivity scores. Therefore, this score (% above/below average) should then be used as a multiplier in combination with the average weight produced by a particular seed stock. For example, the bell-curve of the 2145s should be multiplied by the score of the 2145, the bell curve of the 1059.6 V/MG should be multiplied with the score of the 1059.6 V/MG etc., and the results derived from this combined approach should (in my opinion) be a good measure. Anyway, it seems to be hard (close to impossible) to obtain the data for such an analysis.

12/6/2017 4:06:51 AM

Grittyville, WA

Yes, Radach's 1223 stood out.

Personal-bests might occur because partly because of the year it was grown in, partly because it was grown in the best location available to the grower, but overall its probably nearly impossible grow a personal best off a dud seed. I do think growing a personal best (I like to call them ATB's all time bests) indicates something special about that seed...

Here's my story this year: I didn't grow any really big ones, I did grow a lot though, around 50 plants, and its surprising to see how much a seed that has the right stuff can outperform a seed that doesn't. Overall most they fall into a curve where there are many good plants, some less good plants, and a few complete duds. For some of the plants I grew, I feel like I can say "that seed had the right stuff... in anyone else's patch it probably would have performed very well...(maybe 1200 - 1600 lbs?)"...

Also there was one unexpected outlier in the 50 plants I grew. Maybe it for any seed, for every fifty plants grown there is just gonna be some freak that weighs 25% more. Is that just genetics alone or does that kind of a result require multiple things converging?

Maybe the 559 lb freak from my patch, and the 1223 lb freak in Radach's, and the 2363 freak in Hollands... maybe the genetics of these were all equally exceptional. Although the sizes ended up being different, what they have in common is they were all 20-30% above the others in their peer group.

Which seed will have an outlier that will weigh in at very nearly 3,000 lbs?

Why am I writing about this when what I really care about is color and shape!

12/6/2017 5:34:36 AM

Grittyville, WA

@ pumpking I will attempt to use the bell curves and the "best of patch" (year best) and all-time-best information to rank every seed. (Unless someone wants to save me some time and beats me to it or I'll go ahead anyhow and we can compare results...) Yes the pumpkins that are commonly grown for orange / howard dill get smashed in the statistics and could be use a handicap. But back to baseball, some of these orange pumpkins... its more than needing a handicap its more like pulling a random guy out of the stands and telling him to put a jersey on. There's just zero useful data really. There's data, yes, but the data is so bad its practically like there is no data.

12/6/2017 5:57:32 AM

Streator ,Illinois

I guess what I am trying to say is compare sample size, a seed planted 4 times and a seed planted 100 times is not a good comparison . to much of a range, I would think you would have to narrow the range to compare seeds . 4-10 , 11-20,21-50,51-100 and then above 100, to get a better look at a seeds potential. and then narrow it down to 1st, 2nd ,3rd year growing season. I think it would tell a lot more about a 1st year seed and so on. then a person could use his own variables to judge a seed for himself. variables being grower novice, hobby, or in it to win it. weather heat, cold ,hail or what ever a person wants to use. but for now I like not following the herd , I like the idea of planting new seeds and guessing.

Apologies for being a bit owly my wife says I git that way when on the mid night shift.
and yes she slapped me upside the head to set me right!

as a side note can someone tell me when you ask 10 people you get 11-12 answers the same
and 3 ball players and only get to choose from 2 :)
to answer wizzy another riddle --- grower/seed.
Gene

12/6/2017 6:25:32 AM

Grittyville, WA

I think the shape of the bell curve for the 50 different seeds I grew, from various seeds, would have almost perfectly matched the shape of the bell curve for the progeny from a single seed, grown from by many different growers.

The shape of the bell curve is either determined by 1) the various environments the plants were grown in 2) variable genetics within a single lineage approximates the variable genetics from many closely related lineages 3) both

I wonder if the bell curve from genetic variability matches the bell curve from environmental variability. I doubt it matches perfectly. It sounds funny I find these bell curves slightly mesmerizing. The (half) bell shape says something and means something but you aren't sure what it means, its like looking at the stars in the sky. What am I high on!!!

Does the height of the curve does vary 1) only because of the genetics 2) only because of who grows the seed or the amount of care and expense they put into growing it 3) a mix

Probably #3 in both cases, no surprise, just wanted to wake up all those sleeping neurons...

12/6/2017 7:06:20 AM

Grittyville, WA

Info overload? Yes.

12/6/2017 7:11:35 AM

Willamette Valley, Oregon

Pumpking - In case you missed it, I did do a best-in-patch analysis. I didn't show all the seeds because I really don't put much faith in it as a useful metric. But it does discredit the idea that the 2145 was the greatest BECAUSE it was the best in so many patches. 2145 could still be the greatest seed ever, but that metric ain't why. See my post on: 11/30/2017 6:30:56 PM
It doesn't say, but that's for 2017 only.

12/6/2017 12:20:54 PM

Germany

bathabitat - Your post from 11/30/2017 was one of the things which made me start thinking about those things a bit more. Yes, the 1685 Scherber produced a large percentage of best of patch pumpkins. Now it would be interesting to see by how many % the performance of this seed was better than the others in that patch, and finally this percentage shouldn´t be the measure itself, it should be used as a multiplier in order to upgrade the bell curve of the seed stock.

12/6/2017 12:34:31 PM

Willamette Valley, Oregon

Gene, yes the sample size does effect the reliability of the ratios. Grown a few times with "1 big winner" (is it real or good luck?) in a metric will put that seed higher on the list, than grown several times with "ONLY 1 winner" (is it real or bad luck?) when if they were grown 100 times they'd both be in the middle. That does work against the 2145; because it was grown so much it gets closer to the average. My point in those lists is not to focus on the rank, but to say there are many seeds statistically indistinguishable from the 2145 by many metrics.

Here's what I have for the number weighed for each seed this year (2017) after standardizing seed names the best I could (I suspect there could be a more than a few errors lurking). You can see it quickly drops off to fewer than 10 grown per seed. A >10 fruit cut-off is pretty restrictive; only 20 seeds make the cut with 186 fruit among them, compared to nearly 2000 AG and squash combined. You'll also note the 2145 is without peer, with 3x the second most grown seed.

TotaltoScale Seed YearofSeed
76 2145.5 McMullen 2015
24 2095.5 Willemijns 2016
23 1844.5* Holub 2016
21 2261.5 Wallace 2016
16 1975 Wallace 2015
14 2066 Geddes 2016
13 2624.6 Willemijns 2016
13 1800.3 Hoelke 2016
and... Number of Seeds x Number of times that seed reached the scale
(e.g. 12 seeds were grown 4 times, 11 seeds grown 3 times, etc.)
12x4, 11x3, 10x4, 9x3, 8x3, 7x6, 6x10, 5x7, 4x22,
And the ones I typically ommitted from the analyses because chance plays too large of a role: 3x32, 2x90, 1xthe rest.

But I digress, ha ha, I ultimately think that Gene's (and Cornhusk's) earlier sentiment that all seed assessments are BS is closer to the right answer than anyone cares to admit. What a boring conclusion, though! ha.

12/6/2017 1:01:24 PM

Gays Mills, Wisconsin

bathabitat wrote:
"Cornhusk. I have to point out that your final argument invalidates your metric inasmuch as it invalidates mine."

You are correct, that's the point of my post.

12/6/2017 5:50:46 PM

JACKSON, WISCONSIN. ; )

My brain hurts.... send me all the seeds from above and I’ll grow the best 18 plants and we shall see in fall I know 2000 won’t happen with that many plants but we should find the best ones of the ones grown that way ?

12/6/2017 6:11:15 PM

stoughton WI

Let's all agree to disagree, and let's grow some big fruit. That's what we're all here for anyway. So somebody find that new great seed because there won't be many 2145's left soon. Good luck to all. And gary you probably already have all those seeds lol.

12/6/2017 6:57:19 PM

nOVA sCOTIA

Not only that but there will never be another grower in this hobby as generous as Gene McMullen !!! I am so sick and tired of hearing about low seed counts !!! UP Goes the price

12/6/2017 9:34:26 PM

Eastern Pennsylvania

My brain has turned to sludge after all of this. In the words of the HD king Norm Gansert, “orange rules.” Life is much simpler this way. Have fun and don’t forget to plant your own seed. The biggest pumpkin in your patch will be the one grown from the seed and subsequent plant that you believe in the most.

12/6/2017 10:29:31 PM

Willamette Valley, Oregon

Cornhusk. Now that you point it out, I hear the sarcasm. ;)

Every post on this thread has been with a smile on my face, so I hope no one is taking it any other way. It's December, what else do I have to do?

12/7/2017 1:04:51 AM

Germany

Usually it starts that way (let´s call it "December talk"), but sometimes it can result in something really new, really useful...it´s the crazy ideas which have the greatest capability of bringing about something new.

I like this thread.

12/7/2017 7:14:46 AM

Littlerock, WA

Here is how everyone can be happy: Plant a 2002 Tobeck! You get a 2145 selfed from a perfect plant, steady growing 2k pumpkin.....AND....we can all make Scott happy by marching forward with new genetics. It's a win-win!

12/7/2017 2:32:27 PM

Cape Breton Nova Scotia

I'm going to try and win a couple 2002 Tobeck's at the club auctions. I've got a feeling this seed will see lots of dirt in 2018

12/7/2017 4:34:14 PM

Zimmerman, Minnesota

This is a side question.....as me being a new grower I think I follow what your say but......if Gene had grown on that same plant a pumpkin say 1000 pounds less but with the genetics that were identical...would his theoretical 1145 be able to grow a 2624....or does it need the size (+2000 pounds) to show that the genetics are of a potential to hit a world record. Or is it that no one would try an "1145" so the genetics are never tested.....or maybe this question...what is the smallest pumpkin to produce a 2000+ pounder?

12/7/2017 8:03:10 PM

stoughton WI

answer to the last part of your question. the smallest pumpkin technically to grow a 2000lb pumpkin is the the 1338 martin, it grew a couple over 2000lbs. i was curious so i had to check.

12/7/2017 8:18:38 PM

Streator ,Illinois

If the 2145 was a 1145 I still would have every seed and the squirrels would of loved them but if the same plant produced a 2145 and 1145 was pollinated by same male flower it could in my opinion be the closest thing to a twin you could get but it is a guess .

I believe it is 1338 martin seed

Gene

12/7/2017 8:46:48 PM

Gays Mills, Wisconsin

Hey Farmer Brown,
Excellent questions! Let's pretend Gene picked his 2145 early for some reason and it only weighed XXX pounds. The seeds would be exactly the same and if Mathias got the exact same seed from the XXX it would still be able to grow his 2624! But it is unlikely that Mathias would have even considered Gene's XXX McMullen '15.

12/7/2017 9:28:18 PM

Willamette Valley, Oregon

There is a correlation between seed weight (i.e. mother-fruit weight) and offspring-fruit weight. Here's an old graph, but the same is true in the 2017 data:

http://www.bigpumpkins.com/displayphoto.asp?pid=6883&gid=-50353

The question is - do good growers pick bigger seeds and grow big fruit or do the bigger seeds make better growers? Probably both, but we can't say for sure.

What you can say for sure is that big mother seeds have shown that they have the genes to go big. Small fruit, such as genetic patches, MIGHT have great genetics, but they might not (see clone cross exception). Small-fruit genetics are undetermined and thus lower probability. Every seed is different, but families should have some similarities. We need to make selections to achieve genetic advancement.

Even better than a demonstrably BIG mother is a BIG mother AND a BIG father. See the post at 11/24/2017 10:27:41 PM.

Small fruit from clone crosses (like the 282 Scherber (1725 Harp [2009 Wallace] x 1725 Harp [2009 Wallace] self) can be small but still have verified goodness in their genes (100% 2009 Wallace-fruit genes! - I still love that seed!). Small fruit from seed x clone could also be good, with 50% verified genetic value. Small fruit from seed x seed might be good, but where there's a choice, picking large fruit x large fruit has higher probability of A) possessing the genes needed for bigness and B) passing those genes to the offspring.

12/8/2017 1:01:43 AM

JACKSON, WISCONSIN. ; )

Gene did pick his kin early.... if he would have held out till say Stillwater how much bigger would it have been..... ?
But I know gene loves the cedar burg weigh off and I’m glad he did what he did..... when I rolled up to the hotel and seen that monster kin holy cow was I shocked

12/8/2017 9:59:25 AM