Patrick A. Peterson, PhD

1.1K posts

Patrick A. Peterson, PhD banner
Patrick A. Peterson, PhD

Patrick A. Peterson, PhD

@ppeterson_

The Real Patrick Peterson. Numbers guy. @RFootball #CHOP 🪓

Katılım Mart 2018
2.4K Takip Edilen608 Takipçiler
Patrick A. Peterson, PhD
Patrick A. Peterson, PhD@ppeterson_·
It has absolutely been questioned, the original data comes from Counsilman’s paper from the 80’s where he looked at his swimmers subjective feelings of speed power etc in events and then their training logs prior to it was and always had been conjecture. No control groups nothing. There are actually more recent detraining studies in soccer and rugby athletes that I’ve come across where they go 2-3 weeks with nothing and no significant declines in MSS some showing improvements. People have taken this chart as written law or truth when in reality Counsilman himself proposed it as a framework, basically a principle of his training plans for a non-locomotor sport. Way overblown!
English
1
0
9
809
Jake Tuura
Jake Tuura@jake_tuura·
You lose maximal speed after 5+-3 days. Why is this never questioned? Why do we just trust the 70s/80s Russians?
Jake Tuura tweet media
English
12
12
164
26.5K
Patrick A. Peterson, PhD
Patrick A. Peterson, PhD@ppeterson_·
Hey man I really love your content, been reading for some time now. Excited to see what you guys do at DropBack... I am pretty sure the model you've created here is actually a piecewise (segmented) regression. Stepwise regression is for feature selection (i.e. adding or removing covariates based on AIC). I definitely agree with the theory here but the write up might be a little confusing since the results are a univariate model and stepwise regression is used for big models (I use it a lot with recruiting data). Some more interesting stuff you can find with piecewise (or segmented as its commonly called) is what type of anthropometric measures actually matter and at what levels of other variables (i.e. arm length doesn't matter as much at 6'5"+ because your wingspan with your shoulder/torso is already so wide). Heres another example on piecewise (segmented) regression using NFL data: journals.lww.com/nsca-jscr/full…
English
0
0
0
15
parker fleming
parker fleming@statsowar·
One way to think about how to interpret speeed/measurables outside of tape is with a stepwise regression. Give credit based on certain thresholds, and specifically penalize players who don't make the thresholds. Player goes from 4.5 to 4.7? -35 percentage points From 4.4 to 4.3? +5 percentage points
parker fleming tweet mediaparker fleming tweet media
parker fleming@statsowar

Jotted down some words about interpreting athleticism in player evaluation. Written in the context of the combine, but applicable to college recruiting/portal certainly.

English
1
1
19
8.5K
Patrick A. Peterson, PhD retweetledi
Big 🐰
Big 🐰@FreddieGibbs·
25 a beast
English
7
12
233
63.7K
Patrick A. Peterson, PhD retweetledi
Sharran Srivatsaa
Sharran Srivatsaa@sharran·
A pattern I’ve noticed: People who build tiny things daily become unstoppable within months. People who wait for big moments to take big actions stay exactly where they are for years.
English
144
2.1K
16.9K
374.7K
Patrick A. Peterson, PhD retweetledi
Jash Dholani
Jash Dholani@oldbooksguy·
Nietzsche wrote in 1889 that overworking is a "modern vice." You get lost in a thousand little things that don't matter so you have an excuse to avoid the 2-3 big things that are heavy, difficult, and important. Work becomes a tool not for power, or even happiness, but evasion...
English
51
1.3K
13.6K
422.1K
Patrick A. Peterson, PhD retweetledi
SMB Attorney
SMB Attorney@SMB_Attorney·
I think about this all the time
SMB Attorney tweet media
English
249
4.1K
65.7K
1.9M
Patrick A. Peterson, PhD retweetledi
NBC Sports
NBC Sports@NBCSports·
OH, WHAT A (K)NIGHT 🤩 Rutgers put on a SHOW. 🤌
English
7
41
371
34.1K
Patrick A. Peterson, PhD retweetledi
NFL Draft Files
NFL Draft Files@NFL_DF·
Rookie CBs ranked by passer rating allowed so far in preseason: #Chargers Eric Rogers (0.0) #Bears Jeremiah Walker (39.6) #Chargers Nikko Reed (39.8) #Texans Jaylon Smith (46.3) #Ravens Marquise Robinson (50.3) #Vikings Zemaiah Vaughn (51.0) #Titans Marcus Harris (57.1) #Ravens Keyon Martin (66.0) #Patriots Kobee Minor (69.7) #Bucs Jacob Parrish (70.5) … #Broncos Jahdae Barron (79.2) #Saints Quincy Riley (91.7) #Chiefs Nohl Williams (100.0) #Cardinals Denzel Burke (110.0)
NFL Draft Files tweet media
English
6
34
132
23.7K
Patrick A. Peterson, PhD retweetledi
VraserX e/acc
VraserX e/acc@VraserX·
GPT-5 just casually did new mathematics. Sebastien Bubeck gave it an open problem from convex optimization, something humans had only partially solved. GPT-5-Pro sat down, reasoned for 17 minutes, and produced a correct proof improving the known bound from 1/L all the way to 1.5/L. This wasn’t in the paper. It wasn’t online. It wasn’t memorized. It was new math. Verified by Bubeck himself. Humans later closed the gap at 1.75/L, but GPT-5 independently advanced the frontier. A machine just contributed original research-level mathematics. If you’re not completely stunned by this, you’re not paying attention. We’ve officially entered the era where AI isn’t just learning math, it’s creating it. @sama @OpenAI @kevinweil @gdb @markchen90
VraserX e/acc tweet media
Sebastien Bubeck@SebastienBubeck

Claim: gpt-5-pro can prove new interesting mathematics. Proof: I took a convex optimization paper with a clean open problem in it and asked gpt-5-pro to work on it. It proved a better bound than what is in the paper, and I checked the proof it's correct. Details below.

English
961
3.3K
24.8K
4.9M
Patrick A. Peterson, PhD retweetledi
Dr. Dominic Ng
Dr. Dominic Ng@DrDominicNg·
Your brain ages at the same speed whether you have a PhD or high school diploma. Your daily choices - exercise, sleep, social connection - matter more than the degrees on your wall.
Dr. Dominic Ng tweet media
English
23
131
909
37.5K
Patrick A. Peterson, PhD retweetledi
Tyler Burch
Tyler Burch@TylerJBurch·
it's kind of wild how many billions of dollars are affected by XGBoost models, and yet somehow the current version is 3.0.4 and CRAN is still on 1.7.11.1
English
1
1
15
3.2K