In this post, I will explore and present hypotheses on how the Shrine in Monster Rancher 2 determines the statistics of the monsters it reads from CDs.

Prologue:

My main goals are to show evidence that:
• The attacks, form and nature, likes and dislikes are not determined from the CD subcode data (other than that which determines the actual breed), but the statistics (Power, Life, etc.) are.
• Each stat for each monster generated at the Shrine falls into a narrow range, and the Shrine does not create monsters with wildly unusual values for a stat (except rares like Sueki Suezo).

I would also like to open up these topics for exploration:
• precisely what subcode data is read to determine a monster's stats;
• whether a CD (or subcode data) can be created to offer maximum values for each stat in a particular breed;
• whether a rare monster's stats can be altered through subcode modification; and ultimately
• the possibility of engineering a CD (or subcode data) to create a custom monster (with specific breed as well as stats).

Background:

This research began as a side project when I was trying to crack the Shrine program to be able to create a CD containing a specific breed of monster. (The original project has been reduced to an exercise in curiosity and research for research's sake due to the discovery and popularity of CloneCD, but I'm still looking into it.)

Through my original research, I had created a number of Frozen Gaboos on CD-Rs by altering the track lengths of a two-track CD I had created. I noticed that all the Frozen Gaboos had the same form, nature and attacks, but different stats, likes and dislikes. I also noticed that each stat had a similar value for each Frozen Gaboo, falling within a 50-point range at most.

I thought it would be worth looking into this further, seeing whether it was just a coincidence based on my small sample size (at that point, five) or whether there was a pattern, and whether it could be exploited to create specific monsters with specific stats.

Method:

I opened my personal MR2 CD database and looked for monsters for which I had multiple CDs. I also included the CD-Rs I had burned in my previous research. I discovered I had 10 CDs that would generate Frozen Gaboo [#291, Gaboo/Tiger], 11 CDs that would generate Melon Suezo [#250, Suezo/Zuum], and 9 CDs that would generate KnightMocchi [#275, Mocchi/Durahan]. I created the monsters from each of these CDs at the Shrine, noted the statistics, form, nature, attacks, likes and dislikes.

I had previously burned a reference disk (two audio tracks from .wav files, each under two minutes, using Adaptec's Easy CD Creator). I then burned an exact copy of it (by burning the same .wav files to another disc). Each CD created the same Frozen Gaboo with the same stats, form, nature, and attacks, but different likes and dislikes.

I used the second CD at the Shrine again, and received a Frozen Gaboo which was identical to the first two, except its likes and dislikes were different. Based on this, as well as various threads on this board, it became apparent that likes and dislikes were generated randomly at the Shrine, but that monster types, stats, form, nature and attacks were not random and were based on either the CD data or specific to the monster breed itself.

Many more CD-Rs were burned during this research. Since most of the findings aren't directly related to my goals in this post, I'll save them for another post when I get more data. And much of this information is nothing new to some or most of you, as much of this research has been done before. But to summarize my findings there (and these are still hypotheses):

• The actual data contained in the tracks is not used to generate anything at the shrine. [Changing the volume of the .wav files, as well as making them entirely blank but preserving the length, did not alter any stats or other characteristics.] Only the subcode data is used.
• Likes and dislikes are random and not connected to CD subcode data.
• Nature, form and attacks appear to be specific to the monster's breed and are generated based on the breed, not any other CD subcode data.
• For completeness' sake, I'll also note that the monster's breed is generated based on the CD subcode data.
• The six statistics are based on CD subcode data. [This is evidenced by the fact that the same CD, or identical CDs, consistently produce the same statistics.]

This said, I looked at the data for the statistics of the 30 monsters I had generated (this data is shown below) and noticed certain trends.

Data:

Note: CDs used have been assigned a number (based on the card number of the monster they produce) and a letter (sequential in order of testing). Although the specific CD information is not vital to the data, a list has been included at the end of this post for reference.

CDs which create Frozen Gaboo [#291, Gaboo/Tiger]
(all CDs created monsters with normal form, good nature and chop/slap attacks)

CD	Lif	Pow	Int	Ski	Spd	Def	Comp
291-A	177	117	66	101	161	68	690
291-B	190	116	48	100	169	81	704
291-C	184	118	80	114	159	61	716
291-D	167	126	62	100	185	47	687
291-E	193	135	53	131	153	48	713
291-F	150	154	90	100	175	40	709
291-G	169	135	61	108	170	67	710
291-H	177	127	79	120	159	46	708
291-I	160	131	62	109	185	65	712
291-J	167	153	66	119	186	40	731
min	150	116	48	100	153	40	687
max	193	154	90	131	186	81	731
diff	43	38	42	31	33	41	44

CDs which create Melon Suezo [#250, Suezo/Zuum]
(all CDs created monsters with normal form, bad nature and spit/tail assault attacks)

CD	Lif	Pow	Int	Ski	Spd	Def	Comp
250-A	92	133	162	134	121	109	751
250-B	78	145	153	149	109	115	749
250-C	103	118	176	146	106	99	748
250-D	103	121	168	156	93	101	742
250-E	91	125	166	147	117	118	764
250-F	98	141	161	148	107	104	759
250-G	101	123	156	140	121	113	754
250-H	97	127	172	150	112	111	769
250-I	91	125	166	147	117	118	764
250-J	73	144	152	160	118	98	745
250-K	90	139	162	140	114	112	757
min	73	118	152	134	93	98	742
max	103	145	176	160	121	118	769
diff	30	27	24	26	28	20	27

CDs which create KnightMocchi [#275, Mocchi/Durahan]
(all CDs created monsters with normal form, bad nature and head butt/slap attacks)

CD	Lif	Pow	Int	Ski	Spd	Def	Comp
275-A	110	125	145	158	146	149	833
275-B	112	123	132	154	161	149	831
275-C	96	136	143	172	157	136	840
275-D	106	115	121	165	159	160	826
275-E	114	112	139	167	148	147	827
275-F	102	132	134	153	142	152	815
275-G	111	115	136	167	157	158	844
275-H	104	114	137	165	159	144	823
275-I	94	118	143	176	141	162	834
min	94	112	121	153	141	136	815
max	114	136	145	176	161	162	844
diff	20	24	24	23	20	26	29

Analysis:

When collected as indicated above, the data almost speaks for itself on some points. The widest range for a single stat of a single monster type is 43 (Life for Frozen Gaboo ranges from 150 to 193), and the narrowest range is 20 (Defense for Melon Suezo, Life and Speed for KnightMocchi).

Most surprising to me was the fact that the composites of the stats also fell into a narrow range, with the widest of the three being for the Frozen Gaboo. For 10 monsters (admittedly a small sample size, but still significant), the composites ranged from 687 to 731, a 44-point range. I had expected that, if the stats were somewhat randomly determined, the composite range would reflect some unusually strong and weak monsters in each set.

Since I haven't looked specifically at the subcode data itself, I haven't been able to draw any conclusions on exactly where the numbers for the statistics come from. Here are some hypotheses on how the stats could be generated.

Since the numbers fall into narrow ranges within each stat for each monster, but vary considerably as a whole when comparing different stats for different monsters (one Frozen Gaboo has a Defense of 40; another Frozen Gaboo has a Life of 193), I imagine there's a table of fixed values somewhere in the game to which "random" numbers taken from the subcode data are added.

For instance, since Frozen Gaboos from my sample range from 150 to 193 in Life, this could indicate that the Life stat for a Frozen Gaboo created at the Shrine could be determined by adding a fixed value (possibly 150) to a "random" number from the CD subcode data, modified to fit into a range from, say, 0 to 49, so that all Frozen Gaboos will have Life stats of 150 to 199.

Likewise, the Power stat for a KnightMocchi could be 110 plus a "random" number from the CD subcode data, modified to fit into a range from 0 to 29, so that all Knight Mocchis will have Power stats of 110 to 139 (which would be in keeping with the range of 112 to 136 I found in my nine samples.

Note: When I say a "random" number from the CD subcode data, this is shorthand for saying a number derived from a fixed piece of data from the subcode, which, due to the varying contents of CDs, will create a seemingly random number. It is not truly random in the sense that dice rolls are random, because the same CD will produce the same number each time the Shrine reads it.

And when I speak of modifying the number, there's a mathematical term called modular arithmetic ("mod" for short) that can be used to do this. Using modular arithmetic, for example, using a modulus of 50, 173 becomes 23 because 173 divided by 50 is 3, remainder 23. [173 mod 50 = 23] This could be used to modify a number from the subcode (perhaps a two-digit hexadecimal number, ranging from 00 [0] to FF [255]) to fit into a specific range.

Now a table with entries for each of the 408 monster breeds (or even the 391 playable breeds) with base values for each stat would have over 2,000 entries, which seems bulky programmingwise. I imagine it's far more likely that each of the 38 breeds has a base value for each stat, and that the value for a specific monster type is determined by combining the values for the main and sub breeds, with the sub reduced in importance. (This is how lifespans work, according to other people's research.)

So perhaps the base Life number is 120 for a Gaboo and 60 for a Tiger, and the program adds 120 plus half of 60 to get 150 for a Gaboo/Tiger. If a Joker's base Life number is 40, then 140 (120+(40/2)) would be the base number for a Gaboo/Joker. And a Gaboo/Gaboo would get 180 (120+(120/2)). These are purely hypothetical numbers for illustrative purposes; I haven't tried (nor do I have enough data yet to attempt) mapping out a table of these numbers.

Conclusion:

Once again addressing the points in the prologue:

• The attacks, form and nature, likes and dislikes are not determined from the CD subcode data (other than that which determines the actual breed), but the statistics (Power, Life, etc.) are.

My data supports this hypothesis.

• Each stat for each monster generated at the Shrine falls into a narrow range, and the Shrine does not create monsters with wildly unusual values for a stat (except rares like Sueki Suezo).

My data supports this hypothesis.

• precisely what subcode data is read to determine a monster's stats

I have not had a chance to explore this yet.

• whether a CD (or subcode data) can be created to offer maximum values for each stat in a particular breed

From my data and analysis, it appears that the Shrine program limits the stats generated by using some formula. If this is correct, it would be impossible to create a CD containing a monster with maxed-out stats or even a single stat much greater than the average monster of its kind.

• whether a rare monster's stats can be altered through subcode modification

I have not had a chance to explore this yet. However, since rares (or monsters which have a ??? sub, more specifically) appear to have preset statistics (since only one CD hypothetically produces each monster), it may be impossible to alter the stats. This depends whether the Shrine program, for example, sees that the CD produces a Kasumi (Pixie/???), stops there and gives it a preset set of statistics or not.

• the possibility of engineering a CD (or subcode data) to create a custom monster (with specific breed as well as stats)

I have not had a chance to explore this yet. I am fairly sure it can be done; I just don't know specifically how.

Epilogue:

This is all sheer hypothesis. My sample sizes were quite small for scientific research. I haven't tried comparing data between monsters with the same main breed and different sub breeds. And I haven't even begun to look at the subcode data itself to see where the numbers might come from.

I welcome all comments, criticism, praise and questions, either through this message board or through e-mail. I especially would be interested in data which contradicts my findings. And congratulations for reading the entire post. =]

- Quincunx


Appendix: CDs used to generate data

250-A: The Complete Encyclopedia of Games, disc 1 (PC data)
250-B: Diablo II, cinematics disc (PC data)
250-C: Who Wants to Be a Millionaire promotional CD from General Mills (PC data)
250-D: PaRappa the Rapper (PSX)
250-E: Personal MP3 archive: 1980s.mp3 disc A (PC data)
250-F: Personal MP3 archive: klassiks.mp3 (PC data)
250-G: Quicken Basic 98 (PC data)
250-H: Microsoft Windows 95 (PC data)
250-I: Personal MP3 archive: Napster v1.0 r.i.p. (PC data)
250-J: The Sims Livin' Large Expansion Pack (PC data)
250-K: Wheel of Fortune (PC data)

275-A: Personal MP3 archive: 1980s.mp3 disc B (PC data)
275-B: Diablo II, play disc (PC data)
275-C: Microsoft Windows 98 Upgrade (PC data)
275-D: Que Using HTML Second Edition (PC data)
275-E: Personal MP3 archive: Trance etc. (PC data)
275-F: Final Fantasy VII, disc 1 (PSX)
275-G: Hot Shots Golf (PSX)
275-H: Revelations: Persona (PSX)
275-I: Various Artists: Big Hard Disk vol. 2 (audio)

291-A: Various Artists: Serious Road Trip, disc 2 (audio)
291-B: Various Artists: Chill Out Phase Two, disc 1 (audio)
291-C: Various Artists: Black Box (Wax Trax compilation), disc 3 (audio)
291-D: Various Artists: California Dreaming (audio)
291-E: Aphex Twin: Analogue Bubblebath 4 (audio)
291-F: CD-R #01, burned during research (audio)
291-G: CD-R #14, burned during research (audio)
291-H: CD-R #15, burned during research (audio)
291-I: CD-R #16, burned during research (audio)
291-J: CD-R #17, burned during research (audio)

Great Job, Quincunx. This seems to explain why we don't find the occasional super-monster off CDs. And modular arithmetic is a common gimmic in database programming. People have used it for years to keep casual users from crashing out systems. (Thanks for the term BTW. Used the trick for years without knowing it had a name.)
You might want to check in the MR2 archive under Miscellaneous Q&A for Lisa's thread 'That Crazy Monol'. It had every stat at 25. The subcode from that might tell you which entry to look at for stat generation

I posted that Monol in the MR2 Hard-To-Find section. Also, you might want to look at a thread in the MR2 section, Misc Q & A for a thread about the best possible stats. Several of us posted high and low numbers found on actual CDs there.

John: Thank you! And you're welcome re: the term "modular arithmetic." I've used the trick a lot in programming, but it had been a while since I learned the name for it, and I had to hit the dictionary to remember the word "modulus."

And I checked out Lisa's Monol. Most bizarre! I'm wondering if this is a sort of unique monster, like a ??? sub breed, only given a normal sub breed, but still assigned stats (like Sueki Suezo's).

As Lisa says, "We don't know if this is a function of subcode settings, or if this CD was programmed into the special chart used to make the /??? monsters, and is simply a joke!" My guess is the latter. Maybe someone on the programming team didn't like Def Leppard. =]

Lisa: Thank you! I glanced at the subcode for "that crazy Monol"... and it just shows me how little I know about actual subcode. I'm assuming the first few Entries (Entry 0, Entry 1, Entry 2) don't correspond to audio tracks? Because Entry 3 starts at 0:02.32, and the rest of the Entries progress as tracks would, a few minutes apart. But 0, 1 and 2, if they were interpreted in the same way, would start at 1:00.00, 15:00.00 and 70:32.47.

*striking forehead with palm* Oh... could the 1:00.00 indicate 1 session on the CD, the 15:00.00 indicate 15 audio tracks, and the 70:32.47 be the total burn time (length of CD)?

I *really* need to study up on subcode.

And I have read the thread on best possible stats. (I've actually read the better part of all the MR2 archive messages.) That's one of the things that led me to believe my hypotheses might be right.

you go, quincunx!

i am probably the most inept person on the board w/r/t this sort of research and analysis, yet i find it fascinating. i look forward to *anything* you want to do in this regard.

Yeah, your second part is correct, the CD has 15 audio tracks, total time 70:32.47.

Torey: Thanks! I guess I've still got a bit of the research bug in me from ages ago when I did some biology research work at an actual lab. (Maybe I'm overcompensating for some of the poor work I did there. «grin»)

Thanks for your support!

Lisa: Glad I figured that out... I guess this might be my first step on the road to understanding the subcode.

By the way, although "that crazy Monol" is obviously interesting for other reasons, I don't think it will help me in my current research because I suspect this is a monster built into the game (complete with predetermined stats); I don't imagine the six 25s are read from the subcode.

(shudders to think) If they were read from the subcode, that would be a strong indication that the entire Shrine algorithm was based around a Def Leppard greatest hits CD. (still shuddering) (doesn't want to admit he sang along with the radio in the '80s to more of the songs on that CD than he has fingers on one hand) (and played air guitar) (and air drums)

But seriously, thanks for the info.

All this sounds very possible. I noticed certain monster having a base for low stats when starting off but I never looked into high stats. The monsters at the market have the base low stats if that helps any. I figured out mostly through combining baby monsters over and over again what the lower end of a monsters stats were. It was just something I was doing for fun though and I never wrote down any results, I do remember that pure joker's intelligence at its lowest was 200. I'll try to look into it some this weekend.

mepersoner: Interesting concept, combining baby monsters to get minimum stats. I may have to look into that myself. I'll have a bit of time this week to explore more as to how subcode works and how it determines the stats.

Actually, what I need to do badly is to sit down and play some serious MR2. «grin» I only just today unlocked Beaclon and Mock, and I still have nine more breeds to go.

Maybe we can compare notes on this later. After all, you know what they say about two heads...

Gah! No! Wait! What am I saying? *blink*

Combining baby monsters isn't going to give me the minimum stats that can be generated from a CD at the Shrine. And that's what I'm after. Although I'm sure the minimum possible stats for a monster breed would be interesting and useful, I'm specifically looking at monsters straight off the CD.

Or maybe I'm jumping the gun here. I guess it depends on this: Is it possible to breed two monsters with one or more stats below the minimum that can be pulled off a CD? For instance, if Frozen Gaboos straight from CD have a Def of 40-89, is it possible to combine monsters to form a Frozen Gaboo which starts out with a Def of less than 40?

If this can be done, then I can't cross baby monsters in the Lab in search of minimum stats obtainable at the Shrine.

My eyes are getting very heavy, so I'll worry more about this after I get some sleep...

Oh yeah, for the record, I went out and bought this CD after someone else told us about the stats thing. It's not a CD I would have bought for the music, ever.

Well it seemed (could be wrong, but pretty sure)
that the mininum stats aff of cd is the same as
the mininum through combining, but I'm pretty sure
the maximum is completely different.

Lisa: Nice disclaimer... you mean you weren't singing along with "Pour Some Sugar on Me" every time it came on the radio? (chuckles) I was a child of the '80s, and while I didn't profess my devotion to Def Leppard by scrawling the band's logo on my book covers and jean jacket with a black Sharpie, I still sang along with the songs when they came on the radio... but only when I was alone. «grin»

mepersoner: Thanks for that info. I'll check it out.

Quincunx: Here's some data for your research. I was working on modifying Carole King's "Thoroughbred" to get Trotter from it. Just started but I noticed an oddity. The original CK with PMIN=33 PSEC=20 FRAME=50 produces a Purple Worm (Worm/Naga) with the following stats:

151

113

120

121

51

110

A modified CK with PMIN=33 PSEC=14 FRAME=50 produces a Purple Worm again, but with these stats:

175

112

131

124

49

103

Based on this very unscientific sample PSec seems to have an intimate relationship with stats.

John: Thanks for the data! Is this the Pmin/Psec/Frame for the entire album or for just the entry for one of the last tracks you changed the data for? I'm unfamiliar with the actual album.

I've gotten some indications from old/new data that, at least for the first track, adjusting the length in seconds lets you keep the main breed and collect scattered sub breeds. Adjusting the length of the second track by a few seconds seems to let you keep the main and sub and collect scattered sets of stats. But changing the length of the second track by a full minute caused a change in main and sub breeds.

Note: These results are VERY preliminary, and specific to a very small subset of custom CDs I've burned, each of which only has two tracks. So don't quote me on this, and don't take it as gospel. «grin»

Based on my last two paragraphs, and noting that a Trotter is a Centaur/??? and you've got a Worm, I really don't have any recommendations. Actually, I note that cdnow.com lists the album as 33'11" in length and you're trying lengths of 33'20" and 33'14". (I also see the album is from 1976, so who knows how many pressings there have been of it?)

Again, thanks!

I also had time to try 33'11" before I crashed last night. Just decided to post this morning. This is changing the Entry 2 Psec entry.
Purple Worm

150

139

108

127

52

119

John: I was rather ambiguous in my previous post, I guess. I mentioned the 33'11" because that was what cdnow.com lists as the track length. I don't know where they got that number from. Not only are there possibly multiple pressings, but the number may have been gotten by adding up the track times shown on the CD packaging itself, and these times are notoriously a second or two (or 10) off. What I'm saying is... I don't really know where their 33'11" comes from.

Also... thanks for telling me that you were changing the Entry 2 Psec entry. Based on what I know so far about subcode, this means that, indeed, you've changed the "burn time" for the entire CD in the subcode.

[When I say "burn time," I'm referring to the total length of the album, from the first frame of the first track to the last frame of the last track, as opposed to the time it takes to burn the CD on your CD-R. Am I using the right term here, and if not, would someone please correct me so that I don't confuse anyone else?]

It also means, hypothetically and to the best of my understanding, that you've changed only the length of the last track, and that all other tracks remain their original lengths, in the subcode.

Based on some primitive and preliminary old research, I'd say the MR2 Shrine looks at more than just the total burn time of the CD. When I burned a CD with two tracks of 2'00.47" and 0'46.03", it made a Dokoo [Gaboo/Joker]. When I burned a CD with two tracks of 1'59.47" and 0'47.03", it made a Hare Hound [Tiger/Hare]. Yet the burn time of these two CDs are identical. [Note: The .47 and .03 refer to 1/100ths of a second, not to frames.]

In summary, it may take more than just altering the Entry 2 time to make the rare. Knowing that audio CDs can have one track or several, I wouldn't be surprised if the shrine looked at number of tracks [Entry 1], total burn time [Entry 2] and the length of the first track [Entry 3] to determine the monster.

I had hoped cddb.com would help me. CDDB is an Internet database of CD data that allows a program (like Audiograbber [an MP3/wav ripper] and some computer CD players) to read some data from the CD, go to the database and download the artist, album title and track names. I haven't had a chance to read all of their FAQ, but I couldn't give their site to give me (by accessing it with a browser) the track lengths and total burn time for Carole King's "Thoroughbred"; all it would give me was track names.

This is a shame, because CDDB does have an entry for Carole King's "Thoroughbred," and it does thave that information for at least one pressing of the CD! And if you put a copy of the CD into the drive and used a program that used CDDB, it would confirm that, yes, based on the data, this is Carole King's "Thoroughbred"! But it won't give me the data when I give it the artist and title; maybe they're protecting their proprietary data. «grimace»

Anyway, I've rambled long enough, and I'm late for work again. «grin»

Just a quick note about people listing times for tracks: when I first started looking into the topic I noticed that different CD players would read tracks from the same CD as having different lengths (when you viewed the info on the screen, not anything to do with subcode)(my PSX and stereo gave very different times from each other). This was originally cause for some confusion when Torey and I were exchanging data. It was only when I started using clonecd that I could get accurate info.

And now, some further research on the subject. From what I can tell, the Shrine looks at four spots for stats-Entry 1 PMin, Entry 2 PMin, Entry 2 Psec AND Entry 2 Frame.
Pmin1 defines a range. Modifying Carole King's "Thoroughbred" from 10 down to 1 produced the same Purple Worm with no change in overall stats. Changes up produced an Ape and a Jell.
PMin2 tightens the range. Even small changes here can produce whole new monsters.
PSec2 is the main source of stats. Small changes here change stats up and down in unpredictable ways. The algorithm that Tecmo uses must be really complex. Also, large changes down produce whole new monsters.
Frame2 was just added after further research showed that a 5 frame change down gave our "Mellow Gold" Mum Mew 131 LIF.

Alpha Working Hypothesis
Thus, to tie it all together. I believe that PMin1 and PMin2 define Main and Sub in very general terms across several breeds. Stats are generated from PSec2 and Frame2. Then a 'Best Match' monster is chosen (say four out of six stats in the breed's range) and any stats that are outside the bounds are modified to fit.
It's very raw, but it covers all of the noted phenomena except Lisa's Crazy Monol. Any ideas?

John: Wow! Well-done! I don't agree with everything you've said, but there's some interesting stuff there.

If you'll note the fifth paragraph of my July 4 post in this thread, the two CDs described there would have identical Entry 1 PMin (number of tracks) and Entry 2 PMin, Psec and Frame (total time), but create entirely different monsters.

I'm also a bit suspicious of Frame2 affecting the monster, as I burned three CDs that were identical except for the first track being 1'58.97", 1'59.47" and 1'59.72" (those are hundredths of seconds after the decimal, not frames), and they all created identical monsters with identical stats. Since the other track on each CD was 0'46.03", we're talking total burn times of 2'45.00", 2'45.50" and 2'45.75", which would produce Frame2s of 0, 37 or 38 and 56 or 57. (I haven't actually looked at the CloneCD data; I'm just doing the math.)

So is it possible that your 5 frame change down actually was read as an entire second loss? (There's a 1 in 15 chance.)

But I do think you're on to something here.

And BTW, I still think Lisa's Crazy Monol is a special rare that pretends to be a normal main/sub combination.

But keep it up... maybe we're starting to get somewhere!

Quincunx: The Frame2 assumption comes from work on Enigma's Magnet. Changes as small as ONE Frame in either direction causes a Suezo to appear in MR1.

From this, we can assume that Rares are exact matches to a look-up table for all four numbers (PMin1, PMin2, PSec2 and the lovely Frame2). {I hate this assumption, by the way. Maybe they are only almost exact?, he says pleadingly.}

MR2 adds something to the mix and that's why we get Magnet, but not Moo. I like your idea of a check on the first actual track, but I'm uncertain how to test it.(And out of CD-Rs anyway.)

The 5 Frame Wonder might be a statistical glitch covered by a stat falling out of range.

I'm beginning to think Lisa's Crazy Monol is a sort of joke. It may be the dummy monster the programmers used to check their coding. The American localization team came across it and just tied it to a CD as a joke. No real way of knowing.

One other location in the subcode data that you may wish to investigate could be the TocEntries value under the [Disc] section at the top of the .CCD file.
In Monster Rancher 1 I found that by altering this value I was able to change the main type of the monster created at the shrine.

After reading through some of the more recent research posted here, I concur that the PMin value from [Entry 1] seems to be related to the number of audio tracks on a disc. If you are modifying an audio disc, add 3 to this value and it should match the TocEntries value.
I had overlooked this bit when doing my earlier tests, and it may be a point of interest worthy of further investigation.

Something else that I have observed, and seems to hold true from my limited experience, the [Entry 0] values correspond to the start of the music tracks (or number of sessions, I truly do not know about this one), [Entry 1] contains the number of audio tracks, and [Entry 2] has the total length of audio data on the disc. The Entries after this are the individual tracks on the disc and their corresponding starting locations.

Another value you may wish to play with I would recommend a couple tests with the PLBA value from [Entry 2]. This is a calculated value equal to the total number of frames on the disc (PLBA = PFrame + 75 * (PSec + 60 * PMin) - 150 ). I did not see any affect in Monster Rancher 1 when altering this upwards from what it should be, but I do not know if it has an affect in Monster Rancher 2. I found that by changing it to be too low the CD would not read at the shrine. Just to be safe, I would recommend modifying this to correspond to your new PMin, PSec and PFrame values in [Entry 2].

I hope to be able to assist in further research as my limited free time allows. Feel free to browse my previously posted research under the thread "The quest for ..." for more information.

Best of luck and happy ranching.

John: You had me baffled with the reference to Magnet, but now I realize it's a MR1 rare. And, yes, I would believe that the Shrine uses Frame2 (as well as PMin1, PMin2 and PSec2) to verify that a disc contains a rare.

My assertion that Frame2 doesn't change a monster was intended only to apply to non-rares. That's why I thought maybe the subtraction of five frames might have pushed the PSec2 down by 1. This would match up with your belief that PSec2 is used to determine stats.

And kudos on the comment on Lisa's Crazy Monol! That could be it exactly. It's either there as a dummy monster, a joke or both. (I chuckle to think of two programmers sharing an office, and one programmer constantly playing that Def Leppard CD, driving the other programmer to make darned well sure that that CD would produce a lousy monster.) «grin»

Enigma: I'm honored to see one of my predecessors in research dignify my post with a response. Thank you.

You mentioned editing the number of TocEntries. I'm not quite sure what exactly would do to the subcode. A 10-track CD would have entries 0 through 12. If you reduced the TocEntries by 2, would it just throw out the last two entries? Perhaps it wouldn't even burn the subcode for the last two entries to the disc.

The PMin value in Entry 0 seems more likely to correspond to the number of sessions on the disc than the start of the music tracks. Otherwise, there would be a minute of blank space burned at the beginning of the CD, no? I notice that when I'm burning a music CD on my burner, it takes 1'00" or 1'01" to create the TOC. However, that's at 4x speed; it takes 2'00 to 2'02" at 2x speed, so I guess it might be four minutes' worth of CD space used for the TOC data.

Regarding the PLBA value (and TocEntries, for that matter), I guess it depends whether the shrine is reading these values or the actual PMin, PSec, etc. It would seem that changing one without the other might have no effect.

And I'm quite familiar with your "The quest for..." thread. Your work and CatsGodot's work were the main reasons I became interested in doing my own research. I notice most of the data in the "The quest for..." thread is for MR1, though, and while I'm sure the MR1 and MR2 shrine programs are quite similar, I note that I'm working exclusively on MR2 research.

Thank you for your information!

I think it would help greatly if someone could produce more than one CD subcode for the same rare monster.

If someone has the Zebrahead CD with Zebrasaurian on it, we can calculate the subcode against the one I found on a CD of mine (House of Groove - posted in the subcode data section). That would prove, or dismiss, the fact that rares come from specific subcode data.

I was also told at one point that Ninja Kato could be found on 5 CDs, based on Yoda's partially incorrect listing, which I have a copy of. If anyone has these (and it checks out), maybe we could match it against the Ninja Kato listed here already.

Final Fantasy 8, Disc 2 (Jap/PSX)
*Tenchu Stealth Assassin (PSX)* listed in the Lab already
Close Combat: A Bridge Too Far (PC)
ADAM: The Inside Story (Special Edition) (PC)
March Madness ’99 (PSX)

I have both Bust A Move 4 and BackDraft, they both make CinderBird. The Subcode is identical. This 'ideticality' first noticed in MRBC ep.II, that led me to look into the topic.

WHT Junior: I personally am not researching for rares, but many here are. Thanks for your post!

Ninja Katos on five different CDs? Three PSX discs and two PC CD-ROMs all with the same subcode data? This seems incredible... I'd love to get my hands on the subcode data for these CDs and compare them.

Sadly, I don't have any of these CDs.

Quincunx: I doubt that it would matter if the last two entries were thrown out by changing the TOC. Most PS games have only entry 0 thru entry 3. Our generation information must be in something from Disc down to Entry 3.

Also, I believe Entry 0 is the disc info to indentify the type of disc. All the audio discs I've seen so far have 4350 for the PLBA and all the data discs have 6750. This may allow the game to use a second modulas to bring PMin1 down to a working range as on data discs this can get very large.

Do you want the code for my two Cinderbird discs?

Just so you know, every disc that BattleCard epII saw as identical, that I owned, turned out to be 100% identical code. I printed them out, back when I first started looking at the problem.

Sure, Lisa. Post the Bust A Move code or e-mail it. But really, you know, I've just had a thought.

We should be comparing some of these subcodes we have the rares for against their non-working cousins. There must be one thing besides the stats calculations that determines rares. Otherwise, rares wouldn't be so rare. : )

oh, has anyone else noticed that the PLBA for Entry 2 agrees with what CloneCD identifies as the total number of sectors on the disc?

My only real thought is, aside from the search for Moo, the discovery of what makes the rares has very little to do with deriving a chart of what makes the regular breeds & stats.

It tells us what not to use. That would, I think anyway, be a one use number. Only if it matches do you find the rare in MR2.

'Course, I have been known to be wrong. ; D

Ok, it's gonna take 4 hours for me to run BAM4 so be patient!

John: You do have a valid point there. If the shrine can read a CD with one audio track or a PSX disc, both of which have entries 0 to 3 only, then it seems the Shrine wouldn't use any data beyond Entry 3.

I haven't even been considering data discs, and I don't think I'm going to start quite yet. But I agree with your thoughts on the modulus.

And I must say that I agree with Lisa; analyzing rares (and the regulars found on a different pressing of the same CD) is a discrete endeavor from tabulating what makes breeds and stats.

Lisa: I'd gladly take a copy of the code for the two Cinderbird discs as well. My e-mail address is linked to my name at the top of this post, as always.

And as I mentioned to John, I agree with you that there are two discrete endeavors here: understanding what makes a rare rare, and understanding what makes a monster's breed and stats. With the advent of CloneCD, I have little interest in the former and am fascinated by the latter.

Please note: it's not worthwhile yet for me to split this section of the board, so try and put info in the title of your new posts as to wheter you are doing rare research or research into regular monsters/stats.

But on one experiment, I adjusted the values in the LAST entry of the subcode data, and got interesting results. I got every main for Mew except Pure and Rare. The same happened when I modified the last Entry for Nitons. I got all of the Nitons minus Pure and Rare. IIRC, that Entry was 14.

ok, ok, my memory is extremely faulty....the code is different, and there's no way i looked at them at all (cinderbirds) because I hadn't figured out how to get my drive to read PSX discs...

Ai-yi-yi! Lisa, I don't understand! These two CDs have nothing in common except for their Entry 2's PMin and PSec. I assume, based on the thematic trend in CDs that produce rare monsters, that the Cinderbird's intended rare is the "Backdraft" soundtrack, and that Bust a Move 4 happened to have enough numbers in common with it to also make a Cinderbird. But why?

Surely there are countless CDs that run 42'56". That's a classic album length from those days when albums were 10 songs, 3 or 4 minutes each with at least one "arty" track that ran long. Heck, even I might have released an album of that length in the mid-'90s.

There's got to be more to it than that, but what?

I sound pretty shook up for someone who professes not to care about researching rares, don't I?