What Is A Robbery? Devin Haney - Vasyl Lomachenko Scored By Unbaised Artificial Intelligence.

[crimsonfalcon07]

Edited for brevity.

And in regard to Loma varying the strength of his punches, if it lands on Haney, even just a touch jab, it is still a scoring blow. It would not be a significant punch, but it is supposed to be factored by the judges. So I'm not sure what you're saying here.

First. Point is I'm about as close to a sympathetic ear you're going to find on here, and I'm telling you for a fact that based on your thread title, this doesn't look like a thread about AI. I've engaged with you in the other threads too, and I'm quite clear on what's going on here. I get that you are fanboying hard on this company. Good for you. Doesn't change anything of my analysis.

It's even more obvious when you resort to name calling and accusations of bias and emotion. That's a textbook logical fallacy.

Second. This comment of yours I've highlighted gives me the impression you don't really understand boxing scoring. You seem to think that the ten point must system can be objective, and that all that matters is that it be a scoring punch. But that's not what the scoring criteria says. Scoring punches are a facet of what the judges consider, but they're not the be-all end-all, because all punches are not created equal. As you said, they're factored by the judges, but that's not the same as a clear definition for how they're supposed to be weighted.

So third, that's the major thing you're missing here that people have tried to tell you repeatedly. I just identified an error for you-that it's not effectively weighting distraction punches vs power punches. It's doing the analysis based on speed, and not doing a good job evaluating it on impact.

You can see this throughout the fight, especially if you're watching to see when the fighter is sitting down on the punch or using another power mechanism to transfer weight into the punch. Just in the first couple rounds I could see mistakes in how it weights the power of punches. For instance, in the first round, it weights a body punch of Haney to 3, but there's very little impact because Loma rotated his body on impact and the punch ends up sliding. Then it rates a short upper from Loma at one, even though that's a direct impact that stands Haney up a bit, and you can see the hips travel forward instead of just being an arm punch.

There's a lot of these things that are really obvious if you've studied mechanisms for power that the AI doesn't appear to consider at all. That's part of why I said it's flawed from the outset. Any algorithm is going to be based on what the human programs into it, and that's a really basic error that's leading to flaws just at its own baseline premise. Then you get into the notion that there's not an objective way to translate the data into a 10 point must, and the entire thing goes out the window. Is it better than the current corrupt mess? Sure, because it's a fan project that's not making huge money from corruption. That's better on the outset. But a trained monkey pushing buttons might also be an improvement, so we're not saying much here.

Another way to improve things would be to have an option for a pool of other judges, assigned randomly among those requesting to participate, to score the fight remotely from multiple camera angles, and those scorecards could be used to vet discrepancies in the main 3, or considered as well as part of the final scores. They could be paid a fixed fee to essentially work from home.

Or there could be an option for fights deemed controversial to be rescored after the fact by a larger pool of judges.

Or there could be official penalties for judges whose scorecards deviate too much.

There's a lot of ways that the current process could be improved with minimal effort. None ever get any traction, and you know why.

[Cypocryphy]

First. Point is I'm about as close to a sympathetic ear you're going to find on here, and I'm telling you for a fact that based on your thread title, this doesn't look like a thread about AI. I've engaged with you in the other threads too, and I'm quite clear on what's going on here. I get that you are fanboying hard on this company. Good for you. Doesn't change anything of my analysis.

It's even more obvious when you resort to name calling and accusations of bias and emotion. That's a textbook logical fallacy.

Second. This comment of yours I've highlighted gives me the impression you don't really understand boxing scoring. You seem to think that the ten point must system can be objective, and that all that matters is that it be a scoring punch. But that's not what the scoring criteria says. Scoring punches are a facet of what the judges consider, but they're not the be-all end-all, because all punches are not created equal. As you said, they're factored by the judges, but that's not the same as a clear definition for how they're supposed to be weighted.

So third, that's the major thing you're missing here that people have tried to tell you repeatedly. I just identified an error for you-that it's not effectively weighting distraction punches vs power punches. It's doing the analysis based on speed, and not doing a good job evaluating it on impact.

You can see this throughout the fight, especially if you're watching to see when the fighter is sitting down on the punch or using another power mechanism to transfer weight into the punch. Just in the first couple rounds I could see mistakes in how it weights the power of punches. For instance, in the first round, it weights a body punch of Haney to 3, but there's very little impact because Loma rotated his body on impact and the punch ends up sliding. Then it rates a short upper from Loma at one, even though that's a direct impact that stands Haney up a bit, and you can see the hips travel forward instead of just being an arm punch.

There's a lot of these things that are really obvious if you've studied mechanisms for power that the AI doesn't appear to consider at all. That's part of why I said it's flawed from the outset. Any algorithm is going to be based on what the human programs into it, and that's a really basic error that's leading to flaws just at its own baseline premise. Then you get into the notion that there's not an objective way to translate the data into a 10 point must, and the entire thing goes out the window. Is it better than the current corrupt mess? Sure, because it's a fan project that's not making huge money from corruption. That's better on the outset. But a trained monkey pushing buttons might also be an improvement, so we're not saying much here.

Another way to improve things would be to have an option for a pool of other judges, assigned randomly among those requesting to participate, to score the fight remotely from multiple camera angles, and those scorecards could be used to vet discrepancies in the main 3, or considered as well as part of the final scores. They could be paid a fixed fee to essentially work from home.

Or there could be an option for fights deemed controversial to be rescored after the fact by a larger pool of judges.

Or there could be official penalties for judges whose scorecards deviate too much.

There's a lot of ways that the current process could be improved with minimal effort. None ever get any traction, and you know why.

I've engaged with you in the other threads too, and I'm quite clear on what's going on here. I get that you are fanboying hard on this company. Good for you. Doesn't change anything of my analysis.

It's even more obvious when you resort to name calling and accusations of bias and emotion. That's a textbook logical fallacy.

​

Logical fallacy? So you are trying to act smart now? ... What logical fallacy? When did I call you a name? I see you calling me a fanboy and whatnot, but I don't recall calling you any, so what's got you all emotional with me? Genuinely curious.

Scoring a fight is subjective with objective criteria. However, based on your argument, it's completely subjective, and if it's completely subjective, then why score a fight at all? If you can't quantify and qualify what is happening in a fight, then why bother scoring it at all, if it's so subjective? The truth is that you can actually see how many punches someone lands and who lands the harder punches. That is actually quantifiable. Now I'm not saying that this AI is perfect, but it is far better than what is being done presently, and I'm exited about it. But you call that fanboying? That's really weird.

Second. This comment of yours I've highlighted gives me the impression you don't really understand boxing scoring. You seem to think that the ten point must system can be objective, and that all that matters is that it be a scoring punch. But that's not what the scoring criteria says. Scoring punches are a facet of what the judges consider, but they're not the be-all end-all, because all punches are not created equal. As you said, they're factored by the judges, but that's not the same as a clear definition for how they're supposed to be weighted.

I think you misunderstand how to score a fight, so I will illuminate you.

You score a fight based four criteria as follows: Clean punches, Effective Aggression, Ring Generalship, and Defense. With regard to clean punches, you evaluate the quantity and quality of the punch, whereby the more "effective" punches are given more weight in scoring a fight. Furthermore, you need to pay attention to the order of the criteria. Clean punches come first because that's the most important factor in scoring a fight. Then comes effective aggression, but how do you determine "effective aggression"? You are effective but landing clean punches. You can be aggressive and charge forward, but if you are not landing clean punches, then you are not being "effective." Additionally, if you are the "ring general," that means you are making the other guy fight your fight, and by making him fight your fight, you are making him vulnerable to your "clean punches." And with defense, you are avoiding "clean punches," which, of course, ultimately ties back to the first criterium, that being "clean punches."

So, ultimately, everything revolves around who's landing more clean punches. And before you get all uppity again, this is just not my opinion but the opinion of official judges, such as Steve Weisfeld.

I just wanted to lay that out there because it became blatantly obvious that you don't understand how to score a fight, so I hope you found this enlightening.

You can see this throughout the fight, especially if you're watching to see when the fighter is sitting down on the punch or using another power mechanism to transfer weight into the punch. Just in the first couple rounds I could see mistakes in how it weights the power of punches. For instance, in the first round, it weights a body punch of Haney to 3, but there's very little impact because Loma rotated his body on impact and the punch ends up sliding. Then it rates a short upper from Loma at one, even though that's a direct impact that stands Haney up a bit, and you can see the hips travel forward instead of just being an arm punch.

You're going to have to time stamp that because that will help. I'm not sure what moment you're talking about. Furthermore, there are so many assumptions and presumptions in this paragraph that I don't even know where to begin. I never said this was perfect, and the developers did not either. Nevertheless, it's better than what is going on presently. But please do timestamp this "instance in the first round." I want to see what you're talking about.

Another way to improve things would be to have an option for a pool of other judges, assigned randomly among those requesting to participate, to score the fight remotely from multiple camera angles, and those scorecards could be used to vet discrepancies in the main 3, or considered as well as part of the final scores. They could be paid a fixed fee to essentially work from home.
​

Well, that's an idea. But I think you don't realize the fact that DeepStrike doesn't score fights. It's a tool to assist judges in scoring fights because, as of right now, it does not take the data it collects and transmit it into a a final score, so there will still have to be human judges at this point.

Before I write anymore, however, I want to see what it is that you are talking about with the example you provided. Please use timestamps.

(I will reply later because I have to go now, but I will get back to this.)

[TintaBoricua]

A judge is able to award a 10-8 round without a knockdown, if the fighter scoring 10 dominated the round by a huge margin (this is usually only used if a round is total one-way traffic, or perhaps if a fighter was saved from a knockdown because the round expired).â In the case of the 11th round, it was one way traffic the whole round, so an argument can be made for a 10-8 round. It's possible. Depends on the judge.

I know a judge CAN award a 10-8 round WITHOUT a knockdown. They're rare, but they happen. That's why I mentioned all my examples in my post.

When I think of 10-8 rounds WITHOUT a knockdown, I think of rounds where the beatdown was so horrendous that the losing fighter was on the verge of getting stopped. All the examples I mentioned at least one judge scored it 10-8 without a knockdown.

Pacquiao-De la Hoya round 7: De la Hoya takes about 3 or 4 four punch combinations from Pacquiao in the corner and for a solid minute or more it looks like Weeks is going to wave it off. NOTE: Oscar hung in there for one more round before retiring on his stool.

Taylor-Serrano round 5: Serrano had Taylor STIFF LEGGED for nearly a full minute and a half and holding on for dear life. It is a miracle she didn't get stopped or get knocked down. Her bloody expression tells it all. One judge awarded it 10-8 if I'm not mistaken. Somehow Taylor recomposed herself in the remainder of the fight but everybody that watched that live talked about it the next day.

Mayweather-Gatti round 6: By far one of the PRIME examples of target practice with a young Floyd beating on Gatti like a punching bag. Body punches that sound like *******s, four repeated lead straight right hands that snap Gatti's head back. A leaping left hook, jab, left hook combination that all of which land flush. These aren't slow powerful shots like George Foreman. These are all whiplash-inducing, sweat-flying, breakneck speed punches that sound ABSOLUTELY atrocious. Gatti at the end of round 6 looked like Cotto at the end of the Pacquiao fight. Buddy McGirt stopped it in between rounds.

Maidana-Khan round 10: Khan didn't see the looping overhand right and as a result had his legs prancing around like a newborn calf for nearly 2 minutes as he tried to save face and raise his arms as Pacquiao would whenever he wanted to show bravado. By the end of the round, people applauded Khan for his toughness and making it out of the round because of THAT...he REALLY REALLY got beaten up. Khan went on to win a decision.

Thurman-Lopez round 7: Bruh...that left hook by Lopez turned Thurman's head around and had him on his bicycle and STILL he was taking repeated power shots while skirting the ropes, holding his hands out to get space and getting his clock cleaned. I don't like Thurman too much because he talks bigger than his resume and his activity but the amount of punishment he took here without going down stamps his authenticity as a warrior with heart.

All these rounds were one sided traffic. Loma-Haney round 11 was definitely one of the easiest rounds to score in that fight, but I cannot see how it is anywhere near any of the aforementioned rounds. I see that as a clear 10-9 Loma. 10-8 is really stretching it.

[Cypocryphy]

I know a judge CAN award a 10-8 round WITHOUT a knockdown. They're rare, but they happen. That's why I mentioned all my examples in my post.

When I think of 10-8 rounds WITHOUT a knockdown, I think of rounds where the beatdown was so horrendous that the losing fighter was on the verge of getting stopped. All the examples I mentioned at least one judge scored it 10-8 without a knockdown.

Yeah. I agree. I think that's a good way of assessing a 10-8 round. I think when the other fighter is visible hurt and trying to survive, you might have to give a 10-8 round. Personally, I didn't give Loma a 10-8 round in the 11th, but based on what the criteria is for a 10-8 round, that being one-way traffic the whole round, some might award Loma with a 10-8 round. I mentioned it because someone else mentioned it as a possibility. I can't remember who the trainer was at the time, but it was someone in the industry.

I don't know if you remember, but one judge (oh yeah, Moretti) gave Tank - Garcia a 10-10 round because, in his opinion, Ryan was so dominant that it negated the knockdown by Tank. So apparently one judge's threshold is so loose that he'll give someone who was damaged and injured a 10-8 round, which is actually interesting because he's the same judge that didn't even give Loma the round were he was dominant. (And this is why these judges are a problem because they don't even adhere to their own standards.)

Pacquiao-De la Hoya round 7: De la Hoya takes about 3 or 4 four punch combinations from Pacquiao in the corner and for a solid minute or more it looks like Weeks is going to wave it off. NOTE: Oscar hung in there for one more round before retiring on his stool.

Yeah! Perfect example. I think that should be the standard for a 10-8 round. I remember that. I think this is what we are thinking, where someone is visible hurt and incapable of returning fire.

Taylor-Serrano round 5: Serrano had Taylor STIFF LEGGED for nearly a full minute and a half and holding on for dear life. It is a miracle she didn't get stopped or get knocked down. Her bloody expression tells it all. One judge awarded it 10-8 if I'm not mistaken. Somehow Taylor recomposed herself in the remainder of the fight but everybody that watched that live talked about it the next day.

Well, if that's the case, then I'd imagine you'd give Loma a 10-8 round because it was similar. Haney was hurt and began clinching too. (Or maybe not, I'm trying to remember what Taylor looked like in that round. I'd have to go back and check, but I remember a similar scenario. I recall Taylor was actually throwing back and landing, if I'm not mistaken, so maybe that wouldn't qualify so much.)

Mayweather-Gatti round 6: By far one of the PRIME examples of target practice with a young Floyd beating on Gatti like a punching bag. Body punches that sound like *******s, four repeated lead straight right hands that snap Gatti's head back. A leaping left hook, jab, left hook combination that all of which land flush. These aren't slow powerful shots like George Foreman. These are all whiplash-inducing, sweat-flying, breakneck speed punches that sound ABSOLUTELY atrocious. Gatti at the end of round 6 looked like Cotto at the end of the Pacquiao fight. Buddy McGirt stopped it in between rounds.

Yup. Another good example, similar to Pacman and De La Hoya.

Maidana-Khan round 10: Khan didn't see the looping overhand right and as a result had his legs prancing around like a newborn calf for nearly 2 minutes as he tried to save face and raise his arms as Pacquiao would whenever he wanted to show bravado. By the end of the round, people applauded Khan for his toughness and making it out of the round because of THAT...he REALLY REALLY got beaten up. Khan went on to win a decision.

Yes! Exactly. I think these are obvious 10-8 rounds. Good examples.

Thurman-Lopez round 7: Bruh...that left hook by Lopez turned Thurman's head around and had him on his bicycle and STILL he was taking repeated power shots while skirting the ropes, holding his hands out to get space and getting his clock cleaned. I don't like Thurman too much because he talks bigger than his resume and his activity but the amount of punishment he took here without going down stamps his authenticity as a warrior with heart.

I'd have to go back and look at that one. I can't remember too well, but that sounds like a possible 10-8 round too.

All these rounds were one sided traffic. Loma-Haney round 11 was definitely one of the easiest rounds to score in that fight, but I cannot see how it is anywhere near any of the aforementioned rounds. I see that as a clear 10-9 Loma. 10-8 is really stretching it.

Well. Sure. I just mentioned 10-8 round because someone had mentioned that it could have been scored as a 10-8 round in the past, and based on what the criteria is for a 10-8 round, that being where one fighter is bossing the other fighter around the ring, showing complete dominance, and the other fighter is unable to do anything in return, then the judge (in his opinion) can award a 10-8 round.

Honestly, I mentioned it because of some of the strange scoring going on. And I think it's important to note that Moretti deemed Garcia to have been so dominant that round with Tank, that he gave it 10-10. However, Garcia wasn't that dominant, only throwing for a small portion of that round, while Loma was dominant the whole round and didn't get knocked down. Yet, in Moretti's wisdom, not only was that a 10-9 round but the previous round that was similar was given to Haney.

So that just goes to show that these judges are not really scoring these fights based on any criteria but are scoring the fight based on other factors, such as politics, payments, etc. It's become blatantly obvious over the last couple years, with the likes of Rolly getting one bogus decision after another, and other B-side fighters never getting recognized for winning a fight.

It's really gotten out of hand lately, so much so that boxing is taking big hits in viewership. We've seen HBO tank after the Golovkin and Canelo fiasco. We've watched Showtime sink because of low viewership. We have people everyday saying they are done with boxing. We saw many say that after the Loma - Haney fight, such as Boosy and Nelly. And it's not just them, it's a lot of people. In fact, we were getting the MMA crowd to start to watch boxing, but after the Fury fight with Francis, many are saying they aren't going to watch boxing because it's too rigged.

So this is a serious issue in the sport right now. And something needs to change or else the sport will become even more obscure.

Anyway, I like your examples. I think these examples should be used underneath the rule so that judges have a better understanding what a 10-8 round looks like because right now it appears that even the judges don't really know.

[SUBZER0ED]

The key word to consider when talking about the accuracy of Artificial Intelligence is ARTIFICIAL.

[crimsonfalcon07]

Not going to quote this time because that's too much to delete.

I used the term fanboy because you're on here saying "this is the future" and "we should all support this". That's about as textbook fanboy as it gets. No emotion, just calling it as I see it. If you choose to get "emotional" and take it personally, that's on you. I tend to agree with a lot of your points, but my original point here was that you're saying people are clearly being emotional, triggered, etc about the Loma fight, but factually you titled it differently than the others, which is why you're getting different responses. No need to read more into it.

You and I do not disagree that judges are corrupt or incompetent etc. We differ in that you think this is the clear and only solution.

As for the rules bit, I'm very clear on what the rules say. You're essentially doing a copy-paste from what looks like someone else's article about boxing scoring for beginners, but absent anything else. Looks like a DAZN article TBH.

For instance, take the criteria of "clean punching". What does it mean to land cleanly? There's already been some debate about what that means in the other threads. It's not defined objectively. What about effective? You've already provided your own biased definition of that, but that's also not clearly defined by the rules. There's a lot to be said when you know a punch was hard based upon the sound of it and the way the fighter threw the punch, even if the other fighter has a great poker face and is super tough that you just can't tell otherwise. How about aggression? What does that mean? Etc. There's very little in there that's objective. I could debunk every bit of it, but it's fundamentally subjective, and that's for a reason.

As for why score it at all, it's because they wanted to make it look like a sport, and when they just went until someone fell down, it didn't play well with scheduling and fans, and it's hard to make money that way. So they made scores and fixed things to make it look more objective to gull people placing bets. It's never been a "real sport." It's been rife with corruption from the get-go, and it suits it, as a money making business, to have subjective rules that permit rigging matches when they want to. Often they don't need to because the match outcome is predictable based on the matchmaking.

I'm not timestamping it because I'm not wasting any more time. You said you watched it carefully and didn't catch that error; that's through every match. That tells me you don't actually know what I'm looking for when I look at power delivery, and I don't want to spend the time explaining it to you. I'd be more interested if you could come to my gym and glove up and we could go over some of it so you can feel the difference in your own punches, but if you don't see the subtleties, that's fine. It's still a flaw. And I'm not the only one that's pointed that out.

[The D3vil]

You need to watch DeepStrike in action before posting an opinion like this. It's clear to me that you have not actually evaluated DeepStrike. I went out of my way trying to find issues with it today, and it has prove me wrong rather than the other way around.

The technology is ready to be used. Advancements in AI are far more advanced than you seem to realize. And Deep strike accurately assessed the strikes from each boxer in that fight, and it supports what the majority of the world saw that night, that one man was robbed because he was older and a foreigner.

That's just what it was, and that's why we should push for this technology to be used today, right now. I see terrible and tragic decisions made every weekend. I'm tired of it. These boxers/fighters deserve this.

You must be invested in this in someway

It needs to be implemented in lower levels for a few years first.

To just take a new technology & implement it without trying it out at the lower levels yet would be foolish.

Maybe start in the amateur ranks, then club shows, and then finally in about 5 years, we can fully implement it.

[anonymous2.0]

How long does it take this Deepstrike AI to score a fight? Could it do it in real time and could it be checked against a real life punch scorer, like compubox between rounds? I would have to know more before making a decision. Either way I highly doubt the Powers That Be in boxing would approve of this, can't mess up the good thing they got going now.

[Marchegiano]

AI this AI that but youse laugh when the bama computer had Marciano smash all the HWs and pass on title bout predictions to talk to one another instead.



Are youse too ****** to realize those are AI too?

Are youse too ****** to realize AI plugged into the internet is only impressive because youse have failed to use google worth half a ****?


Oh gee oh my, a ****ing program that can search and read its first hit at me so I have to take the exact same steps and retrieve the exact same info BUT this time instead of me reading an entirely generated page I read while it generates the text. woooo that's such an improvement over a dialog tree I don't even know how to handle it and need to go look up other versions of superficial differences I call technological gains.




Y'all ****in' dumb. I can write an AI right now that shares my own bias, dumbass. I can write one that scores exactly like I do, call it non-bias, and promote it like as if it has something to do with the **** being promoted by the media as the future other than the semantics of being able to call it AI.



The ****ing goombas in ****ing Mario is ****ing AI ya dense pricks. Worth a **** AI, ChatGPT, and the AI I wrote in my ****ty Alex Kidd/Megaman crossover game are ****ing worlds apart and what is highlighted here shares more with my punk ass than GPT.


How is it this tech that requires smarty pants of a higher caliber than whole nations can hire popping up in every single corner of the internet? Maybe, gasp, just maybe, these programmers are using the language the programmers who are so elite some entire first world nations have to just wait on what America and China come up with because the talent is bought up are not actually writing anything like what goes on in the elite programming fields but rather use the jargon so that a ****** ****er doesn't know the difference.



Boxing fans score like ****. I can write an AI. You put two and two together ya dumb **** stick.

[dan-b]

Yeah. Here's DeepStrikes Scorecard:

DEEPSTRIKE

	DEVIN HANEY	VASYL LOMACHENKO
1	10	9
2	9	10
3	9	10
4	9	10
5	9	10
6	9	10
7	9	10
8	9	10
9	9	10
10	9	10
11	9	10
12	10	9
TOTAL:	110	118

This just proves how flawed this AI scoring system is. Your title is misleading also. The bias is simply moved to the preferences of those who write the code.