Here are all 6 sets, for ease of comparison, new ones are bolded:

Set 1	P1	P2	P3	P4	P5
1	96	82	67	39	0
2	93	88	58	42	0
3	92	86	51	0	0
4	93	74	63	47	0
5	93	83	50	37	0
Set 2	P1	P2	P3	P4	P5
1	90	84	75	0	0
2	91	80	71	67	41
3	88	80	65	46	0
4	86	81	62	56	0
5	93	78	62	45	48
Set 3	P1	P2	P3	P4	P5
1	90	86	58	58	0
2	96	79	47	0	0
3	90	76	58	63	0
4	74	79	85	48	0
5	93	75	60	55	44
Set 4	P1	P2	P3	P4	P5
1	90	78	75	57	0
2	99	65	47	63	0
3	93	72	71	40	0
4	85	66	90	68	13
5	91	80	46	55	0
Set 5	P1	P2	P3	P4	P5
1	90	84	75	0	0
2	92	71	73	78	28
3	93	83	50	38	0
4	86	81	62	57	0
5	93	78	62	45	48
Set 6	P1	P2	P3	P4	P5
1	96	82	67	39	0
2	88	94	58	42	0
3	88	85	79	0	0
4	87	77	61	64	0
5	88	76	73	43	0

Upon closer inspection this new set (5&6) is really an updated version of 1&2. The original (1&2) came from two seasons ago, and 5&6 is the current rendition.

What is clear is that there was a transcribing error made in the OP.

# 2 of set 1 had his P1 and P2 inverted.

He appears in the newer version as #2 of set 6.

I will correct it in my spreadsheet and try working on it some more. I have a few more learning sets scheduled to come in over the next few weeks.

May be wrong but so far it appears that these "Best Pitches" blurbs are showing up around game 125 or so.
At least that has happened in both worlds I was able to secure data from.

OK another set here:

Set 7	P1	P2	P3	P4	P5
1	94	70	74	69	0
2	94	84	49	57	0
3	95	80	62	56	52
4	86	74	70	72	0
5	90	75	64	58	44
Set 8	P1	P2	P3	P4	P5
1	83	85	83	85	43
2	83	85	72	46	0
3	89	76	64	51	0
4	88	75	68	0	0
5	90	80	49	0	0

So I want to kind of think through this out loud a bit, and welcome other voices.

Look at #3 and #4 in set 8.
89 is one better than 88, 76 is one better than 75, 64 is 4 worse than 68. So either P1 and P2 are so much more heavily weighted that is makes up for the 4 point difference in P3 OR that 51for a P4 is the difference maker.

Let's also look at set 6:

Set 6	P1	P2	P3	P4	P5
2	88	94	58	42	0
3	88	85	79	0	0

Here let's focus on #2 and #3.
88 P1 is a wash, 94 is 9 points better than the 85. But that 58 is 21 behind in the P3. Once again the higher ranked guy has a P4 while the other doesn't. So these are kind of similar to the previous 2 guys.

IF we say well, P1 and P2 have to be much more heavily weighted (and that's exactly what I have in my original formula), THEN we have to look at set 3 #4 & #5:

Set 3	P1	P2	P3	P4	P5
4	74	79	85	48	0
5	93	75	60	55	44

If the previous supposition was correct then how in the world does #4 rank ahead of #5. Nothing about it makes sense going by our line of thinking so far. His P1 is 19 points lower. That's a lot. He is 4 points higher in P2, but that's not enough to make up for the huge difference in P1. Now his P3 is 25 points higher, which is a huge difference, but if P1 and P2 were weighted heavily enough to explain the previous pairs, then this P3 difference wouldn't be enough. Now P4 comes into play and again it is #5 with a 7 point advantage, AND he also has a P5 whereas #4 does not.


Last pair I want to focus on is #2 and #3 from set 4:

Set 4	P1	P2	P3	P4	P5
2	99	65	47	63	0
3	93	72	71	40	0

The difference in P1 is only 6 (advantage to #2), then #3 has a 7 point advantage in P2, a ginormous 24 point edge in P3 and then a similar 23 point edge goes to #2 in P4.

So I think that there are are few pairs of guys in all of these sets that are/should be very telling. Question is what exactly are they telling us?

My guess is that the order of pitches is a difference maker, but a marginal one... and from the data selected there it seems having a fourth pitch might be weighted heavily, but a fifth pitch not so much? I'm going to go through the full sets of data to see if those theories hold true

Actually if you like at num 3 from set 1 it seems like he benefits from not having a fourth pitch like num 5, and he's ahead of num 4 which seems to suggest his better second pitch is more important than his worse third pitch

Posted by groth911 on 5/15/2015 4:04:00 PM (view original):
Actually if you like at num 3 from set 1 it seems like he benefits from not having a fourth pitch like num 5, and he's ahead of num 4 which seems to suggest his better second pitch is more important than his worse third pitch

You would have to assume that if it weren't for the P4 #5 would also be in front of #4. #3 and #5 are really close, enough so that I am fairly confident if all three were 3 pitch guys it would go 3-5-4.

SO that 10 point difference in P4 HAS to be the reason for 4 top be ranked ahead of 5.

I've been looking at this today, and I think I've come to the realization that there is probably data we are missing.  It is saying best pitches.  I'm not sure that this is an absolute - it is likely a relative value.  So it could be weighing each pitch against the average of that pitch for the world.  Which would weigh each pitches value differently in each world.  Just a thought....

my first thought just looking at the numbers made me think they're simply ranked according to average, insinuating that the order of pitches does not matter... here are some totals from your groups:

Another interesting note is that Dave Maduro is not represented in the Mantle set from S34 (Sets 1 and 2), but he ranks 3 in the S36 version.

Now he was in the other league back in S34, which would put him in the group represented by set 2. 

the guy ranked #1 in that set was (at the time) a 3 pitch guy with pitch ratings of 90-84-75. Compare that to Maduro's S34 pitch line of  86-83-77. Certainly not as good, but pretty close. But he  isn't even ranked?

Posted by evegoe on 4/27/2015 11:03:00 AM (view original):

Posted by mchales_army on 4/26/2015 8:10:00 PM (view original):

1	90	86	58	58	0
2	96	79	47	0	0
3	90	76	58	63	0
4	74	79	85	48	0
5	93	75	60	55	44
Not Ranked	92	79	71	57	0

So even though I finally have a formula that will rank all 4 sets in the right order, I went back to the world where the new learning sets came from and found a guy who would be ranked #2 if my formula was applied world wide.

I have included him in the above set with the guys who were all ranked ahead of him. There is NOTHING about his pitches that would seem to make him not ranked above #5 at least.

I wonder what sort of criteria they use to determine who they rank.

Obviously they don't rank any two pitch guys, but I can't see anything that would make sense as to why he wouldn't be included in the ranking.

Here are links to all 6 pitchers from that set:

1) Brady O'Connor
2) Gaylord Enright
3) Ham Joyce
4) Quilvio Cerda
5) Alan McCorley
NR) Matthew Tenbrink

Anything stand out as to why Tenbrink wouldn't qualify for this ranking and the others would?

I was thinking maybe velocity does factor into the list... but then I saw Cerda's rating.
?

1-82 vel	90	86	58	58	0
2-91 vel	96	79	47	0	0
3-81 vel	90	76	58	63	0
4-39 vel	74	79	85	48	0
5-84 vel	93	75	60	55	44
Not Ranked-46 vel	92	79	71	57	0

 

OK. At least this mystery is solved. It seems to only rank those pitchers that have more than 50% of their appearances as a starter.

So even though The one guy ( Fernando Velazquez) was listed as a Long A, he obviously goes back and forth between SP and RP, and at the time of this blurb he had more starts than relief appearances.

If I apply that criteria then it eliminates any pitcher who ranked higher than any of the 5.

Posted by tecwrg on 11/19/2014 8:07:00 PM (view original):
I've tried to use a very simple formula to come up with a weighted overall value for P1-P5 pitches.  It almost works for the two sets of examples you have above.

Basically: ((P1*5)+(P2*4)+(P3*3)+(P4*2)+(P5*1))/(sum of weights for non-zero pitches).

So for instance, the first guy (96,82,67,39,0) would be:

((96*5)+(82*4)+(67*3)+(39*2)+(0*1)) / (5+4+3+2), or (1087 / 14) = 77.6

Using that formula on your two sets of players, we get:

1 - 77.6
2 - 76.8
3 - 79.8
4 - 74.6
5 - 72.9

and

1 - 87.6
2 - 77.5
3 - 74.8
4 - 75.1
5 - 73.4

It's not perfect, as you can see that the 3-pitch pitcher (#3) in the first group is rated the highest, and #3 and #4 is the second group are in the wrong order, but it's close.

But it seems that this formula might be similar in concept to what they do.  Maybe the weights are different for 3 pitch pitchers, 4 pitch pitchers, and 5 pitch pitchers.

Posted by mchales_army on 5/15/2015 6:57:00 PM (view original):

Posted by evegoe on 4/27/2015 11:03:00 AM (view original):

Posted by mchales_army on 4/26/2015 8:10:00 PM (view original):

1	90	86	58	58	0
2	96	79	47	0	0
3	90	76	58	63	0
4	74	79	85	48	0
5	93	75	60	55	44
Not Ranked	92	79	71	57	0

So even though I finally have a formula that will rank all 4 sets in the right order, I went back to the world where the new learning sets came from and found a guy who would be ranked #2 if my formula was applied world wide.

I have included him in the above set with the guys who were all ranked ahead of him. There is NOTHING about his pitches that would seem to make him not ranked above #5 at least.

I wonder what sort of criteria they use to determine who they rank.

Obviously they don't rank any two pitch guys, but I can't see anything that would make sense as to why he wouldn't be included in the ranking.

Here are links to all 6 pitchers from that set:

1) Brady O'Connor
2) Gaylord Enright
3) Ham Joyce
4) Quilvio Cerda
5) Alan McCorley
NR) Matthew Tenbrink

Anything stand out as to why Tenbrink wouldn't qualify for this ranking and the others would?

I was thinking maybe velocity does factor into the list... but then I saw Cerda's rating.
?

1-82 vel	90	86	58	58	0
2-91 vel	96	79	47	0	0
3-81 vel	90	76	58	63	0
4-39 vel	74	79	85	48	0
5-84 vel	93	75	60	55	44
Not Ranked-46 vel	92	79	71	57	0

 

OK. At least this mystery is solved. It seems to only rank those pitchers that have more than 50% of their appearances as a starter.

So even though The one guy ( Fernando Velazquez) was listed as a Long A, he obviously goes back and forth between SP and RP, and at the time of this blurb he had more starts than relief appearances.

If I apply that criteria then it eliminates any pitcher who ranked higher than any of the 5.

Of course, as soon as I post that I find an outlier...

Set 4 

1	90	78	75	57	0
2	99	65	47	63	0
3	93	72	71	40	0
4	85	66	90	68	13
5	91	80	46	55	0
NR	94	78	56	50	47

Posted by pjfoster13 on 5/15/2015 7:04:00 PM (view original):

Posted by tecwrg on 11/19/2014 8:07:00 PM (view original):
I've tried to use a very simple formula to come up with a weighted overall value for P1-P5 pitches.  It almost works for the two sets of examples you have above.

Basically: ((P1*5)+(P2*4)+(P3*3)+(P4*2)+(P5*1))/(sum of weights for non-zero pitches).

So for instance, the first guy (96,82,67,39,0) would be:

((96*5)+(82*4)+(67*3)+(39*2)+(0*1)) / (5+4+3+2), or (1087 / 14) = 77.6

Using that formula on your two sets of players, we get:

1 - 77.6
2 - 76.8
3 - 79.8
4 - 74.6
5 - 72.9

and

1 - 87.6
2 - 77.5
3 - 74.8
4 - 75.1
5 - 73.4

It's not perfect, as you can see that the 3-pitch pitcher (#3) in the first group is rated the highest, and #3 and #4 is the second group are in the wrong order, but it's close.

But it seems that this formula might be similar in concept to what they do.  Maybe the weights are different for 3 pitch pitchers, 4 pitch pitchers, and 5 pitch pitchers.

((96*5)+(82*4)+(67*3)+(39*2)+(0*1)) / (5+4+3+2 [+1]), or (1087 / 14) = 77.6

set 5

83 (3)

68.4 (5)

66 (4) ??

71.5 (4) ??

65.2 (5)

 

450, 336, 225                         1011 /12 84.25     1011 /14 72.214          1011 /15 67.4

460, 284, 219, 156, 28                                                                            1007 /15 67.133

465, 332, 150, 76                                                1023 /14 73.0714        1023 /15 68.2

430, 324, 186, 114                                              1054 /14 75.286          1054 /15 70.266

465, 312, 186, 90, 48                                                                              1101 /15 73.4

doesn't look like that's the formula
 

No it isn't.

My original formula which ranked the first two sets correctly was 8/6/4/1/1.
That worked until the next set came from Centerfield (sets 3&4). In fact set 4 was so off I had to really dig deep.

I finally came up with a couple of new categories which accounted for set 3 #4 having a backwards looking line of 74-79-85, as well as set 4 #2 who had a 99 P1 but the rest of his pitches were rather low compared to every other guy in any of the sets.

When I added both new categories to the 8/6/4/1/1 weight it ordered all 4 sets correctly again. 

Now when the 4 new sets appeared, in set 5 #4 was ranked ahead of 3 but really 3-5 were very very close. There was only a .08% difference between 3 and 5.
However in set 6 #3 was ranked first and #5 was significantly better than #4.

In the next two sets (7&8), set 8 is fine, but set 7 has #3 ranked 1st and #5 ranked 3rd.

The tricky part is that from all of the data I have if I adjust to correct one of the sets it throws the other off even worse...

my head hurts. make it stop!