My prediction for DDP vs Chimaev

Luffy

Banned
Banned
Joined
Nov 2, 2017
Messages
2,793
Reaction score
1,910
Alright, the great Luffy will bless you all with his analysis and visualization skills by sharing my take on how DDP vs Khamzat like goes. I've rewatched their fights already, locked in significant strikes landed per minute, patterns, cardio shown, level of competition... And despite the odds backing the Borz, I think DDP may show why he's one of the best fighters right now. His striking looks awkward but he's a genius. It doesn't seem too technical but it works very well for him — offense and defense. And it's not like his defense is poor, he uses his varied entries to confuse opponents and to defend by attacking. You see that he has a tendency not to let the opponent get a shot in without retaliating immediately, but he does that vs great strikers like against Izzy...

Numbers that matter from their last 5 fights (I won't pick on all their fights otherwise it's a headache) : Dricus lands ~6 sig. Strikes a minute (4th best rate in MW history) and he doesn't fade late. Khamzat's pattern is hell for 5 minutes, then human. Dricus is awkward but doesn't fade specially post nose surgery, and he'll also wrestle if it buys him minutes (6/11 TDs vs Strickland I when DDP was still improving)

My take : DDP beats Chimaev UD (4-1) — 49 - 46 ... Chimaev has a path by an early sub if DDP falters, but unlikely imo as Usman didn't and DDP is tougher than Usman, shown also by the way he body locked Whittaker when he was taken down. Whittaker said he felt a tremendous pressure in the body lock when he took DDP down, adding up to DDP's huge strength. Whittaker and Sean also said they believe DDP wins (Whittaker's opinion matters as having fought both)

Let's visualize the likeliest scenario, which I'll put into writing.


Round 1

The bell tolls!! Chimaev shoots off the first jab and dumps DDP inside the first minute — like vs Usman. He rides half-guard, floats to the back, cranks the face a bit (think the Whittaker finish, just no tap), and keeps Dricus hand-fighting instead of building a base. DDP stands twice but gets mat-returned again like The Wolf generally does. Damage stays light like Usman managed to minimize but the control is heavy. Maybe a 10-8 or 10-9 because I think DDP would escape too much, but a clear domination.

(10-9 Chimaev)


Round 2

Shot's speed from the Borz is already down a note and Dricus starts to stab the body and chop the legs — exactly what he leaned on vs Strickland 2/Izzy when he found his rythm. Clean sprawl mid round, he pummels to doubles, shucks off, cracks a right on exit. He scrambles off a lazy single ending with DDP on top for the first time!! (He said he'd wrestle Khamzat), and DDP drops short elbows. Momentum shifts with Khamzat getting up being being outdone clearly, 10-9 Du Plessis

(10-9 Dricus)

Round 3

Now it looks like Chimaev vs Usman R2-3 again — burts from Khamzat, then breathing. Dricus starts to walk him down, mixing knees in the clinch (Strickland II like template), keg kicks, looping rights. Khamzat goes for a double but is telegraphed, turning into DDP riding half-guard again and landing elbows — he is not just defending, he's stealing grappling minutes, which he does to break Khamzat down psychologically by offering no path.

(10-9 DDP)

Round 4

Khamzat goes for a desperation single, gets stuffed, DDP circles to the back, wrist rides, opens a cut and visible swelling with hammerfists (I think there's a higher chance of a stopppahe here than a round 1 avalanche win for Khamzat). But mostly likely, Khamzat gets back up since he's tough and has a huge heart too. Still, DDP pressures is relentless, and he digs the body. Khamzat clips him once and gets some solid hooks, but in the clinch, DDP explores Khamzat's dipping gas tank and lands knees and ends up with big looping hooks that hurts Khamzat visibly.

(10-9 DDP)

Round 5

Both tired but Dricus' tank holds as he's been there, Khamzat's TD shots are single entries that stall on the fence. DDP knees ribs/thighs, peels off with overhands, sprawls the last TD, spins behind and rides it out with punishing strikes, finishing dominantly.

(10-9 DDP)


49-46 for DDP (second likely to me would be a ~4th round TKO by Dricus, less likely but possible, round 1 finish by Khamzat)

My prediction— "after 5 rounds, all judges score the contest — 49-46 — for the winner, by UD... and Still!!! Dricus Stillknocks Du Plessis!!!! "
 
Alright, the great Luffy will bless you all with his analysis and visualization skills by sharing my take on how DDP vs Khamzat like goes. I've rewatched their fights already, locked in significant strikes landed per minute, patterns, cardio shown, level of competition... And despite the odds backing the Borz, I think DDP may show why he's one of the best fighters right now. His striking looks awkward but he's a genius. It doesn't seem too technical but it works very well for him — offense and defense. And it's not like his defense is poor, he uses his varied entries to confuse opponents and to defend by attacking. You see that he has a tendency not to let the opponent get a shot in without retaliating immediately, but he does that vs great strikers like against Izzy...

Numbers that matter from their last 5 fights (I won't pick on all their fights otherwise it's a headache) : Dricus lands ~6 sig. Strikes a minute (4th best rate in MW history) and he doesn't fade late. Khamzat's pattern is hell for 5 minutes, then human. Dricus is awkward but doesn't fade specially post nose surgery, and he'll also wrestle if it buys him minutes (6/11 TDs vs Strickland I when DDP was still improving)

My take : DDP beats Chimaev UD (4-1) — 49 - 46 ... Chimaev has a path by an early sub if DDP falters, but unlikely imo as Usman didn't and DDP is tougher than Usman, shown also by the way he body locked Whittaker when he was taken down. Whittaker said he felt a tremendous pressure in the body lock when he took DDP down, adding up to DDP's huge strength. Whittaker and Sean also said they believe DDP wins (Whittaker's opinion matters as having fought both)

Let's visualize the likeliest scenario, which I'll put into writing.


Round 1

The bell tolls!! Chimaev shoots off the first jab and dumps DDP inside the first minute — like vs Usman. He rides half-guard, floats to the back, cranks the face a bit (think the Whittaker finish, just no tap), and keeps Dricus hand-fighting instead of building a base. DDP stands twice but gets mat-returned again like The Wolf generally does. Damage stays light like Usman managed to minimize but the control is heavy. Maybe a 10-8 or 10-9 because I think DDP would escape too much, but a clear domination.

(10-9 Chimaev)


Round 2

Shot's speed from the Borz is already down a note and Dricus starts to stab the body and chop the legs — exactly what he leaned on vs Strickland 2/Izzy when he found his rythm. Clean sprawl mid round, he pummels to doubles, shucks off, cracks a right on exit. He scrambles off a lazy single ending with DDP on top for the first time!! (He said he'd wrestle Khamzat), and DDP drops short elbows. Momentum shifts with Khamzat getting up being being outdone clearly, 10-9 Du Plessis

(10-9 Dricus)

Round 3

Now it looks like Chimaev vs Usman R2-3 again — burts from Khamzat, then breathing. Dricus starts to walk him down, mixing knees in the clinch (Strickland II like template), keg kicks, looping rights. Khamzat goes for a double but is telegraphed, turning into DDP riding half-guard again and landing elbows — he is not just defending, he's stealing grappling minutes, which he does to break Khamzat down psychologically by offering no path.

(10-9 DDP)

Round 4

Khamzat goes for a desperation single, gets stuffed, DDP circles to the back, wrist rides, opens a cut and visible swelling with hammerfists (I think there's a higher chance of a stopppahe here than a round 1 avalanche win for Khamzat). But mostly likely, Khamzat gets back up since he's tough and has a huge heart too. Still, DDP pressures is relentless, and he digs the body. Khamzat clips him once and gets some solid hooks, but in the clinch, DDP explores Khamzat's dipping gas tank and lands knees and ends up with big looping hooks that hurts Khamzat visibly.

(10-9 DDP)

Round 5

Both tired but Dricus' tank holds as he's been there, Khamzat's TD shots are single entries that stall on the fence. DDP knees ribs/thighs, peels off with overhands, sprawls the last TD, spins behind and rides it out with punishing strikes, finishing dominantly.

(10-9 DDP)


49-46 for DDP (second likely to me would be a ~4th round TKO by Dricus, less likely but possible, round 1 finish by Khamzat)

My prediction— "after 5 rounds, all judges score the contest — 49-46 — for the winner, by UD... and Still!!! Dricus Stillknocks Du Plessis!!!! "
c92.gif
 
Here we go again, another Luffy gem. What on earth is he gonna tell us this time

source.gif
 
Lol ..... I share how I visualize fights and even make a funny scenario that I imagined for the fight, and that's a stupid thread.... Sure... At least I show why I think what I think, don't say it's a fact and I'm open to new ideas every time
 

I'd definitely have chose another username nowadays. Luffy was a random pick in 2017 since I couldn't think of whatever username as I'm terrible with coming up with new usernames lol. But since I'm known as Luffy by everyone here no way I'm changing that now even tho I wish haha
 
I'd definitely have chose another username nowadays. Luffy was a random pick in 2017 since I couldn't think of whatever username as I'm terrible with coming up with new usernames lol. But since I'm known as Luffy by everyone here no way I'm changing that now even tho I wish haha
Good times though, rewatching those scenes remembers me of my High School days in which I'd come back home and watch like 10 episodes one by one haha good nostalgia
 
Yeah, wording was graphic lol I meant he targets body/legs early to sap gas, not that he’s stalking Camp Crystal Lake

Kudos that you got the reference. I wasn't sure you would.

Good on you, Luff.
 
My early prediction is Dricus via KO/TKO in either round 2 or 3.
 
Well comment what you think of my visualization... Come on, that's how we make threads interesting!!!

Khamzat isn't gassing after one round of him dominating. It's just not happening. If you think he slows noticeably that quickly, I think you're way off. Now after 2 rounds if they're really high paced? You could see him slow. But Dricus does too. The thing is, neither stop fighting even when tired. It may get sloppier (which maybe favors DDP?), but there will still be action.

I think people who believe that "Chimaev is toast if he doesn't finish early" are simplifying things too much. Dricus will show it if he's been mauled for 7, 8, 9 minutes. They'll both be tired but Khamzat will be up 2 rounds.
 
Hey Luffy could you post some of the Python source code for the simulations you've been running?
It’s not magic, I use my yellow lens to sound like a nerd 🤓 and then just NumPy for random draws, pandas to hold the fighter stats, SciPy’s beta/poisson stuff for rates, and a little of a logistic head to turn feature deltas into win probabilities.

I can trim down to share the code here if you are REALLY INTERESTED, if not, then I won't waste my time writing the trim of the code lol...

... Anyways since it's fun and enjoyable, I will write a basics outline ...


-----


import numpy as np
import pandas as pd
from scipy.stats import beta, poisson
from sklearn.linear_model import LogisticRegression

rng = np.random.default_rng(42)




# --- toy data just to show structure ---
fighters = pd.DataFrame([
{"name": "F1", "slpm": 5.6, "sapm": 3.1, "td_acc": 0.38, "td_def": 0.78,
"sub_avg": 0.3, "str_acc": 0.52, "str_def": 0.60},
{"name": "F2", "slpm": 3.9, "sapm": 2.7, "td_acc": 0.42, "td_def": 0.72,
"sub_avg": 0.6, "str_acc": 0.48, "str_def": 0.55},
])

def sample_rate(p, n=200):
"""Sample a probability with uncertainty via Beta."""
# alpha/beta are crude: successes ~ p*n, failures ~(1-p)*n
a, b = max(p*n, 1), max((1-p)*n, 1)
return rng.beta(a, b)

def sim_fight(row1, row2, minutes=25):
# sample noisy versions of key rates
f1_acc = sample_rate(row1.str_acc)
f2_acc = sample_rate(row2.str_acc)
f1_td_a = sample_rate(row1.td_acc)
f2_td_d = sample_rate(row2.td_def)
# expected landed strikes using simple Poisson
f1_exp_str = rng.poisson(f1_acc * row1.slpm * minutes)
f2_exp_str = rng.poisson(f2_acc * row2.slpm * minutes)
# crude takedown math
f1_exp_td = rng.binomial(8, f1_td_a * (1 - row2.td_def))
f2_exp_td = rng.binomial(8, sample_rate(row2.td_acc) * (1 - row1.td_def))
# feature vector for a tiny logistic head
X = np.array([[f1_exp_str - f2_exp_str,
f1_exp_td - f2_exp_td]])
# weights picked by quick CV offline; hard-coded here
w = np.array([0.015, 0.20])
logit = X @ w
prob_f1 = 1 / (1 + np.exp(-logit))
return prob_f1[0]

def run_mc(f1, f2, n=10000):
row1 = fighters.loc[fighters.name == f1].squeeze()
row2 = fighters.loc[fighters.name == f2].squeeze()
probs = [sim_fight(row1, row2) for _ in range(n)]
return np.mean
(probs)

print(run_mc("F1", "F2", n=5000))


----

There are some basic comments so you know what each piece does more or less


It basically samples noisy strike/TD rates with Beta, then turns them into counts (Poisson/Binomial), feeds a tiny 2-feature logistic to spit a win prob. Swapping the actual UFCStats numbers and expanding features (knockdowns, control time, etc.) gives the gist of it....

If you're really interested, I can DM the full outline, but it's big... To sum up, data scrape —> clean —> feature engineering —> CV on historical fights... Showing the whole thing would be too big for a post unless I zip it but since I think you aren't that serious I will leave it at that ....
 
Here's the Chris Benoit breakdown

Round 1

The two meet in the circle and Dricus throws a punch but his thumb goes into Chimaevs eyes. Chimaev goes down hurt. Doctor comes in and wave it off

No contest at 0:37 Round 1
 
Back
Top