Ever since I created this system 4-5 weeks ago which is trained on all UFC events that have occurred in the Paramount era it has gone completely undefeated on its official picks... its pretty low volume on official plays it had 1 official play the first 2 weeks (1 for each card so 2 plays) and then 3 official plays the 3rd week... the plays were Moreno EVEN over Morales in week 1, Van -125 over Pantoja in week 2, Perez -150 over Dumont and Viera -160 over briejzek and Rosas -155 over Barcelos in week 3..... week 4 is unfortunately the first time that it had no plays
Last week (week 4) was about as bad as week 3 was amazing in terms of correctly identifying value on every fight but it had no official plays and thus remains undefeated. So we learned that its leans are too unreliable to be trusted and should be verified by human YT prediction channels. For example I was very skeptical of its edges on Coria, Parkin, Green. Human prediction channels unanimously correctly loved Walker and Imanol and some mentioned Ribovics style should overwhelm Green's hands down style.
I pointed out 4 things before the fights in the ufc 332 thread:
1) Even the system acknowledged that its perceived massive positive value edge on Coria was extremely unreliable with 11% reliability rating so we already knew not to trust this before the fights, easy to go the opposite way with the human predictions
2) It was extremely unsure about Wint <- this proved to be a lack of data problem on both wint and his opponent not a problem with wint himself
3) Correctly identified that Peyton Talbott is so dominant over Deiveson that he was still value even at -650 moneyline
4) Silva looked like an official pick at -200 she just barely didn't qualify numerically
Unfortunately again no official plays 2 weeks in a row I'm gonna have to work on making its official picks higher volume without losing quality. Silva should have been an official pick.
On a quest to be the very best
To remove first post, remove entire topic.
Ever since I created this system 4-5 weeks ago which is trained on all UFC events that have occurred in the Paramount era it has gone completely undefeated on its official picks... its pretty low volume on official plays it had 1 official play the first 2 weeks (1 for each card so 2 plays) and then 3 official plays the 3rd week... the plays were Moreno EVEN over Morales in week 1, Van -125 over Pantoja in week 2, Perez -150 over Dumont and Viera -160 over briejzek and Rosas -155 over Barcelos in week 3..... week 4 is unfortunately the first time that it had no plays
Last week (week 4) was about as bad as week 3 was amazing in terms of correctly identifying value on every fight but it had no official plays and thus remains undefeated. So we learned that its leans are too unreliable to be trusted and should be verified by human YT prediction channels. For example I was very skeptical of its edges on Coria, Parkin, Green. Human prediction channels unanimously correctly loved Walker and Imanol and some mentioned Ribovics style should overwhelm Green's hands down style.
I pointed out 4 things before the fights in the ufc 332 thread:
1) Even the system acknowledged that its perceived massive positive value edge on Coria was extremely unreliable with 11% reliability rating so we already knew not to trust this before the fights, easy to go the opposite way with the human predictions
2) It was extremely unsure about Wint <- this proved to be a lack of data problem on both wint and his opponent not a problem with wint himself
3) Correctly identified that Peyton Talbott is so dominant over Deiveson that he was still value even at -650 moneyline
4) Silva looked like an official pick at -200 she just barely didn't qualify numerically
Unfortunately again no official plays 2 weeks in a row I'm gonna have to work on making its official picks higher volume without losing quality. Silva should have been an official pick.
Hot Forum Topics
