Testing The Accuracy and Predictions of The Computer

cubedrum · Post by **cubedrum** » 03 Sep 2006, 03:33

IrishRufusMurphy wrote:Samuel Peter makes the upset winning a 12 round decision over James Toney, breaking the computers 2 fight win-prediction streak.

Not many of the experts would have predicted Peter winning via decision.

But the computer is right [considering Toney is 233] that Toney could have been undertrained for this fight, as it did give Peter the edge by 8 points.

Toney did manage to win one out of the three judges score cards, but nevertheless the "Nigerian Nightmare" proved the critics wrong, and I should have stuck to my instinctive feeling that Peter would pull the upset.

Nevertheless this computer testing will continue as planned.

Computer is now 2-1 in it's predictions.

I estimated the "undertrained" line for Toney at 230. Computer nailed it. The percentages were close, but given the Toney undertrained factor, the most likely result was a Peter decision. Peter has come in as high as 280 and started his career at 227 so he carries a lot of weight and can still be considered trained.

HomicideHenry · Post by **HomicideHenry** » 03 Sep 2006, 04:27

So, according to Cube, the computer is 3-0... being 100% correct.

All I know is the computer gave us various results on the Peter-Toney fight, and I guess we all should have known Toney was under-trained considering he was 233, and not a natural HW at that. So, I guess the computer was right in ways.

And I did mention Holyfield-Bates before as one of the fights I would run. I have gave my explaination as to why it came in late, and if my explaination isn't to be trusted, then maybe Cube's should, considering he ran the very same fight and got a KO win for Holyfield.

I guess looking at it from Cube's approach the computer hasn't been misleading and has been accurate all three times...I guess the majority expected this computer simulator to judge each man on a 'properly trained' level, but each prediction has been based on numerous factors.

So...

3-0 it is [including the Holyfield 'gimme' with Bates].

OR....for those who don't want to look at it as being right...

2-1 [including the Holyfield 'gimme' with Bates].

Irregardless the 'testing' of this simulator will go on as planned.

Collins2000 · Post by **Collins2000** » 03 Sep 2006, 16:54

I'm bailing out of this fiasco.

If you can't stick to an agreed test plan, what's the point of continuing?

In case you are interested, the final straw was Rupert wildly clutching at lovelarks suggestion that the computer, though its official pick was agreed as being Toney, had actually really picked Peter. With that kind of logic you can prove anything.

cubedrum · Post by **cubedrum** » 03 Sep 2006, 17:19

Collins2000 wrote:I'm bailing out of this fiasco.

If you can't stick to an agreed test plan, what's the point of continuing?

In case you are interested, the final straw was Rupert wildly clutching at lovelarks suggestion that the computer, though its official pick was agreed as being Toney, had actually really picked Peter. With that kind of logic you can prove anything.

When I set up the computer I ran with Toney in shape and out of shape. This is because Toney is a notoriously lazy fat bastard. Then I said that we would see the Toney undertrained scenario if he showed up over 230, which he did. Work on your reading skills dude.

HomicideHenry · Post by **HomicideHenry** » 03 Sep 2006, 18:14

.....well like I said before, for those who can think outside the box and see that the computer was right in the aspect of picking Peter if Toney was undertrained [and at 233 and on basis of his performance] this seems to be the result, it's 3-0.

For those who don't it is 2-1. Mind you, this is the first time this tournament has ever done this before running the same fight over and over 1,000 times and coming up with the percentages.

The first method used for the first two fights were Maskaev/Rahman and Holyfield/Bates, using the 'best two out of three' method that I originally used in the MODERN DAY ATG HEAVYWEIGHT TOURNAMENT.

I only asked Cube/Ezzard and others to lend a helping hand, because from virtually square one I been criticised as 'cheating' the tournaments I been running.

Bottom line is:

1- The computer gave two results, one for both men in condition and with Toney undertrained. The computer hit it on the head considering Peter won a decision [based on figures of Toney being undertrained] as it has predicted.

Surely not every fight in the world that has ever been done has been with fighters who were in 'tip-top' shape. In all fairness having 'prime vs prime' fights should be for 'fantasy fights' like Dempsey vs Louis as an example, and the predictions of an upcoming fight should be based on more factors than 'peak ability', every fighter has an off night.

Anyone who thinks otherwise is a fool. Toney when he first came into the HW division weighed in at roughly 220 when he faced Holyfield, or maybe less and he DOMINATED the former 4x HW champion and Cruiserweight champion, being the 2nd man to knock Holyfield out.

Since then he has been unable to truly dominate a fighter in the way he butchered Holyfield, and WEIGHT is possibly the biggest factor. He weighed 233 for this fight and was 237 when he faced Rahman, both fights were disappointing.

If Toney is to be a player in the HW division he needs to lay off the buffets and hamburgers and lobster feasts and all the booze he can drink, and fight at 215-220 MAX. If he sincerly thinks putting on the pounds is a way for him to increase his punching power, he is dead wrong.

TRAINING makes muscles, not Dunkin Donuts.

Neways....testing will continue as planned....Cube has ran the Valuev/Barrett fight predicting Valuev to win. I will post later my thoughts on this intriguing match-up which will be Valuev's first title defense in America.

Collins2000 · Post by **Collins2000** » 03 Sep 2006, 20:01

cubedrum wrote:
Collins2000 wrote:I'm bailing out of this fiasco.

If you can't stick to an agreed test plan, what's the point of continuing?

In case you are interested, the final straw was Rupert wildly clutching at lovelarks suggestion that the computer, though its official pick was agreed as being Toney, had actually really picked Peter. With that kind of logic you can prove anything.
When I set up the computer I ran with Toney in shape and out of shape. This is because Toney is a notoriously lazy fat bastard. Then I said that we would see the Toney undertrained scenario if he showed up over 230, which he did. Work on your reading skills dude.

My reading skills are fine, mate.

Maybe you ought to to have told us it was really picking Peter, oh I dunno, maybe about the time Rupert was announcing the official computer pick was Toney.

I wonder if you and Rupert would be saying it really picked Peter if Toney had won.

I kind of doubt it.

HomicideHenry · Post by **HomicideHenry** » 03 Sep 2006, 21:38

Hey...I ain't no bear in a cardigan ffs

lol

Neways....the computer did give two predictions... as I have explained this all before in the previous statements above.

pundit · Post by **pundit** » 03 Sep 2006, 22:57

Just shred the computer (and use your brain).

Collins2000 · Post by **Collins2000** » 03 Sep 2006, 23:57

IrishRufusMurphy wrote:Hey...I ain't no bear in a cardigan ffs lol

Neways....the computer did give two predictions... as I have explained this all before in the previous statements above.

Well, it's gonna easily exceed the 70% success rate you are hoping for if you are going to allow it pick both participants in a 2 horse race. Or 1 participant in a one horse race like Holyfield - Master Bates.

In fact, if we could pick both contestants or only put our picks in for 'gimmes' we'd all be at 100%.

Rufus, wasn't this messing with reality the reason why everyone boycotted your heavyweight 'contest'?

HomicideHenry · Post by **HomicideHenry** » 04 Sep 2006, 02:54

Ok. I have gave it some thought and this is what I am going to do:

1- We're going to call the Toney-Peter bout a NC, in all fairness, as it gave two different predictions, though it was accurate dead on on the latter prediction.

2- From now on, no real 'bets' or making a final judgement of the fighters until AFTER the weigh-in, cus in this case [Toney-Peter] the evidence was right in front of us that Toney was over-weight. We need to see exactly who is in shape and whose not to make a good judgement.

3- The scores are still 2-0 [Holyfield gimme and all] with the Toney-Peter bout disregarded, the tournament will go on as planned.

4- No more 'gimmes' or fights that are deemed to be one-sided

Is that fair enough?

Neways...the computer prediction of the Valuev-Barrett fight is based on if both men are trained properly, more or less that both men will be at their best, saying that Valuev will win. If both Valuev and Barrett come in at higher weights than usual...Cube/Ezzard/Myself or someone else should run it as a 'undertrained' simulation...and then go with that.

I hope this settles the peace and brings back the seriousness and 'legit' aspects of this whole testing and project is all about.

cubedrum · Post by **cubedrum** » 04 Sep 2006, 07:03

Collins2000 wrote:
cubedrum wrote:
Collins2000 wrote:I'm bailing out of this fiasco.

If you can't stick to an agreed test plan, what's the point of continuing?

In case you are interested, the final straw was Rupert wildly clutching at lovelarks suggestion that the computer, though its official pick was agreed as being Toney, had actually really picked Peter. With that kind of logic you can prove anything.
When I set up the computer I ran with Toney in shape and out of shape. This is because Toney is a notoriously lazy fat bastard. Then I said that we would see the Toney undertrained scenario if he showed up over 230, which he did. Work on your reading skills dude.
My reading skills are fine, mate.

Maybe you ought to to have told us it was really picking Peter, oh I dunno, maybe about the time Rupert was announcing the official computer pick was Toney.

I wonder if you and Rupert would be saying it really picked Peter if Toney had won.

I kind of doubt it.

You say your reading skills are fine, and in the next sentence completely disregard what I said.

Toney over 230= Peter by decision
Toney under 230= Toney by decision.

That clear enough?

To review: computer 3-0 at picking fights,
Collins 0-2 at reading simple straightforward statements.

HomicideHenry · Post by **HomicideHenry** » 04 Sep 2006, 23:48

Here's my thoughts on Valuev vs Barrett:

1- Valuev is the largest title holder in the history of the ring, his 85" reach and 325 pound 7' frame is a great assett towards him.

2- Valuev has won 'controversial' decisions over both John Ruiz and Larry Donald, his first defense was against #77 ranked Owen Beck. His undefeated record shows great promise but seeing that he is so dominant over grade C fighters yet goes the distance with grade B fighters shows he needs alot of tuning up.

3- Barrett in my opinion will be a genuine good test for the Russian giant, considering he has faced Hasim Rahman and other top 15 men, though he has yet to show the world that he is a top ace himself.

4- It will be a matter of size versus experience against quality opposition. Valuev will outweight Barrett by probably 70 pounds or more, but Barrett has the greater experience against better opposition. The giant has an odd style, sometimes using his size to his advantage but perfers his opponents closer to him. That could be both a good and bad quality to have going against a man like Barrett.

My pick? Barrett failed against Rahman, whom Valuev went the distance and arguably beat John Ruiz who bested Rahman. It is somewhat the even fight, but with an edge to Valuev. My pick is Valuev, despite Barrett having faced better opposition because 'Two Gun' has yet to show me, at least, that he has improved as a fighter---while the giant has improved a bit throwing more combinations.

cubedrum · Post by **cubedrum** » 05 Sep 2006, 18:35

Here's the next test for the computer.
An aging Nelson Dieppa fights Hugo Cazares who defeated him once on Saturday.

Cazares 80.2% KO 48.7%
Dieppa 16.8% KO 8.6%
Draw 3.0%

A walkover for Cazares. Almost 50/50 that Cazares will knock him out. I would have simmed the Chris John fight from that same evening, but his opponent is too obscure to have an official rating.

Ambling Alp · Post by **Ambling Alp** » 05 Sep 2006, 18:37

IrishRufusMurphy wrote:Ok. I have gave it some thought and this is what I am going to do:

1- We're going to call the Toney-Peter bout a NC, in all fairness, as it gave two different predictions, though it was accurate dead on on the latter prediction.

2- From now on, no real 'bets' or making a final judgement of the fighters until AFTER the weigh-in, cus in this case [Toney-Peter] the evidence was right in front of us that Toney was over-weight. We need to see exactly who is in shape and whose not to make a good judgement.

3- The scores are still 2-0 [Holyfield gimme and all] with the Toney-Peter bout disregarded, the tournament will go on as planned.

4- No more 'gimmes' or fights that are deemed to be one-sided

Is that fair enough?

Neways...the computer prediction of the Valuev-Barrett fight is based on if both men are trained properly, more or less that both men will be at their best, saying that Valuev will win. If both Valuev and Barrett come in at higher weights than usual...Cube/Ezzard/Myself or someone else should run it as a 'undertrained' simulation...and then go with that.

I hope this settles the peace and brings back the seriousness and 'legit' aspects of this whole testing and project is all about.

No, calling the prediction on Peter-Toney a No-Contest isn't "fair enough".
Before the fight, there was so much double talk about on who was going to win that I asked on September 2, (2:48) who the computer was picking. You said (just 4 minutes late at 2:52) that the computer had picked Toney.

After the fight on September 3, at 4:40 am you said that it would count as a loss for the computer.

Now you are going back on your word and back to double talk and not counting it as a loss.

Show some integrity and count the result as a loss.
Show some integrity and throw out the Holyfield-Bates since you didn't post this prediction before the fight. (It doesn't matter why you say you didn't. Can't you understand that?)

In the future, just tell us who the computer picks. No if and or buts.

If the is good at picking fights, it will prove it in future fights.

Show some integrity and say that so far the computer is 1 out of 2 in picking fights.

HomicideHenry · Post by **HomicideHenry** » 05 Sep 2006, 19:12

The computer gave 2 predictions for the Toney-Peter fight...yes I did say the computer picked Toney [based on the 'if trained properly factor]. The computer also said if Toney was under-trained, which he was, that Peter would win via decision.

The results of the computer were posted on here for everyone to see, Cube explained the computer would say Toney would lose if he was 230+ pounds, and he did.

NC seems fair to me, but myself, and those who run this program, knows that the computer hit it dead on with the Toney 'under-trained' factor. And like I said before Alp, if people want to view it as being 2-1 thats fine, if they want to view it as being 3-0 that's fine too.

This is part fun and part 'science' and sports...let's leave it as that.

Neways....VALUEV/BARRETT [being judged on peak performance] and the DIEPPA/CAZARES fight have just been 'predicted' by the computer...picking VALUEV and CAZARES to win those fights.

Computer Analysis Thus Far:

MASKAEV vs RAHMAN- '2 out of three' for Maskaev by KO
Actual Result- KO in 12th win for Maskaev

HOLYFIELD vs BATES- '3 out of 3' for Holyfield by KO
Actual Result- KO in 2nd for Holyfield

TONEY vs PETER: Toney to win decision if trained properly; Peter to win decision if Toney trained unproperly
Actual Result- 12 RND win for Peter

AND........I'll tell you what, to make it 'fair' instead of 10 fights to judge, it will be 12 fights to predict, to make up for both Toney and Holyfield. 8)

Can't get any squarer than that.

HomicideHenry · Post by **HomicideHenry** » 06 Sep 2006, 18:08

I think, even though both these fights are scheduled for November, we should run the following two bouts as well:

Floyd Mayweather vs Carlos Baldomir

Shannon Briggs vs Sergei Liakhovich

Both are title matches

Briggs/Liakhovich- WBO title
Maywheather/Baldomir- Welterweight titlist vs Pound-Per-Pound king

I would like to see Oliver McCall-Darroll Wilson run thru as well, cus that match is tonight, so I hope Cube or someone can run it, cus I won't run any of these fights from now on, cus I keep getting accused of trickery.

I just want this wrapped up as quick as we can.

cubedrum · Post by **cubedrum** » 06 Sep 2006, 22:11

OK I simmed the upcoming McCall-Wilson fight.
(based on 1000 fights converted to %)

Oliver McCall - Career stage: End
Training : Emotional Problems
Darroll Wilson - Career stage: End
Training : In Condition

Oliver McCall : 71.7% KO 43.2%
Darroll Wilson : 23.0% KO 5.6%
Draw: 5.3%

McCall gets the nod even though I picked the emotional training condition (one of many options on the sim). McCall is always a toss up how he comes into a fight, so I went with my gut here. Wilson has always had a poor chin, and this combined with McCall's power and granite chin, should make the fight a fairly easy win for the 42-year old McCall.

HomicideHenry · Post by **HomicideHenry** » 07 Sep 2006, 01:17

Sorry lol the McCall-Wilson match happens on SATURDAY...my mistake, thought it was today I misread the article.

cubedrum · Post by **cubedrum** » 10 Sep 2006, 00:51

McCall wins by TKO in 4. Computer was leaning a bit more toward a decision but clearly picked McCall to win.

Computer is 4-0.

HomicideHenry · Post by **HomicideHenry** » 10 Sep 2006, 14:38

COMPUTER ANALYSIS THUS FAR:

OLEG MASKAEV vs HASIM RAHMAN
Computer Prediction- Maskaev by KO
Actual Result- Maskaev KO 12th

EVANDER HOLYFIELD vs JEREMY BATES
Computer Prediction- Holyfield by KO
Actual Result- Holyfield wins KO 2nd

SAMUEL PETER vs JAMES TONEY
Computer Prediction- [Both trained proper] Toney via decision
[James Toney undertrained] Peter via decision
Actual Result- Samuel Peter wins via decision [James Toney 233]

OLIVER MCCALL vs DAROLL WILSON
Computer Prediction- McCall to win
Actual Result- McCall KO's Wilson in 4

************************************************************

I would like to see the following matches be simulated:

Shannon Briggs vs Sergei Liakhovich- WBO title
Floyd Mayweather vs Carlos Baldomir- Welterweight title
Marco Antonio Barrera vs Rocky Juarez
Manny Pacquiao vs Erik Morales
Kevin McBride vs Chris Mallo

This would total up to 12 fights [the Valuev/Barrett fight has been simulated] and alot of them that I just listed are not that far away either, so we can get this project wrapped up.

HomicideHenry · Post by **HomicideHenry** » 11 Sep 2006, 16:18

Also, Cube did a computer simulation of Hugo Cazares vs Nelson Dieppa a while back ago, picking Cazares to win...and that fight is on the 30th, so keep your eyes pealed on that as well.

I would also like to include the fact that most of those fights I listed above to be simulated are happening in a few days [the 16th] and if we're going to simulate them we better do it soon before fight night.

Ambling Alp · Post by **Ambling Alp** » 11 Sep 2006, 21:40

The computer's record is now 2-1.

HomicideHenry · Post by **HomicideHenry** » 12 Sep 2006, 14:02

It isn't 2-1.

I have explained and explained that since it was I who made the goof in saying the computer picked Toney when in fact it gave two predictions [in the case if Toney was 230 and up] so in all actuality it's 4-0.

BUT...since everyone said the Holyfield-Bates fight was a gimme and I screwed up saying the computer picked Toney---instead of being 10 fights to judge it would be 12 to make up for the gimme and the flub I made on Toney-Peter. The Toney/Holyfield bouts are 'NC', based on that fact.

Making the computer 2-0 in predictions, with 6 other fights to be predicted, if Cube/Ezzard can run those [8 fights total] plus the Valuev-Barrett fight that was predicted prior, so that will be 9 fights, with 3 more needed to be added to make the 12.

Ambling Alp · Post by **Ambling Alp** » 12 Sep 2006, 22:46

So if Toney would have won, you would have counted it as a loss for the computer?

HairyBoxe24 · Post by **HairyBoxe24** » 12 Sep 2006, 23:57

Stop bloody having a go at him, this isn't about personal battles.

THE COMPUTER CLEARLY PICKED PETER IF TONEY WAS OVER 230, END OF.
Nothing else matters about who said what, this is about the computer and it clearly picked PETER, now stop having a go about it.

Also i'd like to add although holyfield was a gimme, it even got the round right so tbh its 4-0.