Best 170 pounder ever?

[Claude]

GSP has top 10 ground game in all mma history. Deal with it..

 

[ThunderStruck]

P4P Better on the ground than anyone Fedor beat. Deal with it.

 

[AnotherOldGuy]

How do you know Anderson would have beaten Takase most of the times? Did you watch the fight? That was no fluke. That was a total domination from begining to end. Your statement is completely baseless.

You know who also beat Anderson and had record of 7-4? Ryo Chonan. Again, proving styles >>>> records.

Records are only indicators, but they're not the best indicators to predict a fight. They're better used to assess a fighter's overall success in MMA, but not how that particular fighter matches up against a wrestler or striker. They don't give any insight in any of the variables that happen inside the ring. They don't show strenghts and weaknesses. They just tell who was won and lost more.

Records are quantitative information. Styles are qualitative. When you want to understand how something works, and why it works the way it works, you need qualitative knowledge. Fights aren't numbers. Fights are a bunch of variables meeting each other. To better understand a fight you have to study the those variables because the numbers won't tell you anything about the fight itself.

About team sports, if the higher ranked teams (and thus with better numbers) lose to lower ranked teams, that's pretty much proof numbers are not being reliable. On the other hand if you study variables, suchs the teams' weaknesses, strenghts, injuries, fatigue, etc. you might have a better idea of the outcome. Eg. Germany lost to Mexico in the world cup. It was considered a big upset, as Mexico had never beaten Germany in a world cup. However if you saw Germany in the preparation matches, you would know they were playing terribly, whereas Mexico has always been known as a team that plays better when they're underdogs. To anyone who acknowledged these variables, the result wouldn't have been nearly as surprising compared to the person who only looked at the numbers.

I agree that a lot more than records are needed; as you say, there are a lot more variables involved. Where I disagree is what is the single most indicative indicator.

Suppose we were only give one piece of information about a fighter. You could choose their record (including the organizations in which they fought, as that's part of the record). Or you could choose their style but with no info about their record. Or you could choose their physical attributes, but no info about their record, style, or skill level. Or you could choose their skill level, but no info about their record, style, or physical attributes.

The advantage of the record is that (admittedly to a limited extent) takes the other factors into consideration. A guy with say a 20-0 record in the UFC is going to have pretty decent skills, a fairly flexible style, and at least reasonable physical abilities.

Whereas you could have a guy with a good style matchup against a UFC striker (say a wrestler against Daley back in the day), but who was a weak guy with a 0-20 record in some minor organization. Or a guy with excellent skills, but with a 0-20 record because of physical or mental weaknesses. Or a guy with great physical attributes with a 0-20 because of lack of skill or a horrible style (run the opponent throwing wild punches until exhausted).

Record isn't enough by itself, but it encompasses the other elements better than any other single indicator.

Even when you point out that Takase with 4-7-1 beat Anderson, you're unconsciously using his record to note he fought in Pride, a top level organization. Would you bet on a guy with Takase's style but with a 4-7-1 record in some tiny local organization, against Anderson?


In terms of team sports, you're right that records aren't that useful for single game predictions -- any team (at least in the pro's) can beat any other team on a given day. Where they become useful is when you get a statistically meaningful number of games (baseball in fact, with its 162 game season, because very statistical -- so much so that many fans say its ruining the game, though that's a different topic). That's why many team sports have best of seven series (well, that and because the owners love the play-off fan money). Anyone can win a single game, but the teams with better records tend to win the seven game series.

The problem with soccer (at least World Cup soccer) in that regard is the teams only play each other once. Looking at Germany and Mexico over decades of World Cups gives no useful information, because the teams from even two decades ago have no players in common (and in fact, most teams seem to change a majority of their players over a single decade). How 1960 Mexico did against 1960 Germany tells us nothing about how 2020 Mexico would do about 2020 Germany, they're completely different teams.

Of course, that means record is completely meaningless for World Cup soccer, as you point out. But that's not typical for team sports. In professional soccer (say premier league in England or the Bundesleague in Germany), do teams play each other more often in a season?

 

[Fight Professor]

Obviously Woodley.

 

[Regal Boy]

Serra would beat Usman.

 

[mindyabisness]

GSP 170 Goat

Woodley was great for a while with brutal power but fell off hard. Still GSP for me

 

[Claude]

P4P Better on the ground than anyone Fedor beat. Deal with it.

Especially Nog.

 

[Claude]

I agree that a lot more than records are needed; as you say, there are a lot more variables involved. Where I disagree is what is the single most indicative indicator.

Suppose we were only give one piece of information about a fighter. You could choose their record (including the organizations in which they fought, as that's part of the record). Or you could choose their style but with no info about their record. Or you could choose their physical attributes, but no info about their record, style, or skill level. Or you could choose their skill level, but no info about their record, style, or physical attributes.

The advantage of the record is that (admittedly to a limited extent) takes the other factors into consideration. A guy with say a 20-0 record in the UFC is going to have pretty decent skills, a fairly flexible style, and at least reasonable physical abilities.

Whereas you could have a guy with a good style matchup against a UFC striker (say a wrestler against Daley back in the day), but who was a weak guy with a 0-20 record in some minor organization. Or a guy with excellent skills, but with a 0-20 record because of physical or mental weaknesses. Or a guy with great physical attributes with a 0-20 because of lack of skill or a horrible style (run the opponent throwing wild punches until exhausted).

Record isn't enough by itself, but it encompasses the other elements better than any other single indicator.

Even when you point out that Takase with 4-7-1 beat Anderson, you're unconsciously using his record to note he fought in Pride, a top level organization. Would you bet on a guy with Takase's style but with a 4-7-1 record in some tiny local organization, against Anderson?


In terms of team sports, you're right that records aren't that useful for single game predictions -- any team (at least in the pro's) can beat any other team on a given day. Where they become useful is when you get a statistically meaningful number of games (baseball in fact, with its 162 game season, because very statistical -- so much so that many fans say its ruining the game, though that's a different topic). That's why many team sports have best of seven series (well, that and because the owners love the play-off fan money). Anyone can win a single game, but the teams with better records tend to win the seven game series.

The problem with soccer (at least World Cup soccer) in that regard is the teams only play each other once. Looking at Germany and Mexico over decades of World Cups gives no useful information, because the teams from even two decades ago have no players in common (and in fact, most teams seem to change a majority of their players over a single decade). How 1960 Mexico did against 1960 Germany tells us nothing about how 2020 Mexico would do about 2020 Germany, they're completely different teams.

Of course, that means record is completely meaningless for World Cup soccer, as you point out. But that's not typical for team sports. In professional soccer (say premier league in England or the Bundesleague in Germany), do teams play each other more often in a season?

"A guy with say a 20-0 record in the UFC is going to have pretty decent skills"

That depends. Did he fight all the top tier fighters in the UFC or is he someone like Askren who was 19-0 but most of his wins happened elsewhere? Not every fighter with a good record has been proven in every aspect of the game and under every setting (Unified Rules). Depending on the context, a fighter with a less great record could easily be more proven. That's why records are not optimal predictors.

If there's a 0-20 fighter then you would know he's very low tier just by watching him fight. You could easily see his flaws. You don't need to see CM Punk's record to know he's terrible.

"Even when you point out that Takase with 4-7-1 beat Anderson, you're unconsciously using his record to note he fought in Pride, a top level organization. Would you bet on a guy with Takase's style but with a 4-7-1 record in some tiny local organization, against Anderson?"

I'm stating his record because it's an example that records are poor fight predictors. I would definitely pick Takase to beat Anderson again in Pride at that point in time, regardless of their records. Anderson was just too susceptible to getting outgrappled and that's something his record wouldn't tell you.

Honestly, do you see fight analysts relying on records over styles most of the time or is it the other way around? Like, someone saying "I'm going to pick X fighter because he has a a better record than Y". Or is it a lot more common for them to pick fights based on both fighters' styles and skills? Eg. "I think X guy will win because he has enough TDD to keep the fight standing and superior striking".

I think the answer is obvious.

There's so many scenarios in which records are useless. What if two fighters have the same record? What if they both have decent records? Like say 10-2 and 15-3. What if they have different records, but none of them has a loss? Say 13-0 and 15-0.

You would be 100% clueless because you would have absolutely no insight on the fight, whereas if you knew how the fighters match up, you could easily have a better idea on the outcome.

"In terms of team sports, you're right that records aren't that useful for single game predictions -- any team (at least in the pro's) can beat any other team on a given day. Where they become useful is when you get a statistically meaningful number of games"

If you claim that records aren't useful for single game predictions, then you already admited they're not good for MMA as most MMA fights are single events that usually don't repeat. And if they do repeat, it usually happens after a year, which means fighters momentum has changed and it becomes a single event again.

You just said so: They become useful when you get a statistically meaningful number of games. And that's not MMA by any means.

 

[AnotherOldGuy]

You just said so: They become useful when you get a statistically meaningful number of games. And that's not MMA by any means.

That's actually an excellent point.

Unfortunately style vs style (which exists in team sports as well) is also not a particularly good predictor for a single game or fight, though it probably would be with a statistically meaningful number of games/fights. That's why anyone can win any given game, and why betting on sports isn't a secure way of earning you living, even for people who follow sports closely. People (myself included) tend to be sure we'd know who'd win a fight based on whatever kind of analysis we believe in, but when it comes to actually putting significant amounts of money on the outcome most of us back off -- I've certainly never bet heavily on a fighter based on record or anything else, which tells you all you need to know about how confident I am about my predictions.

So I'll modify my stance based on your points: neither record nor style is a good predictor for a single game/fight -- which is why its so hard to get rich betting on games and fights. If it were simply a matter of looking at records or styles, then we'd all be getting rich by those kinds of analysis -- but in fact, as you alluded to in an earlier post, there are so many variables in MMA (and as I stated in earlier posts, in team sports) that the only people who consistently get rich from sports gambling are the bookies and casinos who take in a percentage of the bets.

Be nice if either way worked consistently though -- a few big bets based on record or style and we could retire very young.

 

[Claude]

That's actually an excellent point.

Unfortunately style vs style (which exists in team sports as well) is also not a particularly good predictor for a single game or fight, though it probably would be with a statistically meaningful number of games/fights. That's why anyone can win any given game, and why betting on sports isn't a secure way of earning you living, even for people who follow sports closely. People (myself included) tend to be sure we'd know who'd win a fight based on whatever kind of analysis we believe in, but when it comes to actually putting significant amounts of money on the outcome most of us back off -- I've certainly never bet heavily on a fighter based on record or anything else, which tells you all you need to know about how confident I am about my predictions.

So I'll modify my stance based on your points: neither record nor style is a good predictor for a single game/fight -- which is why its so hard to get rich betting on games and fights. If it were simply a matter of looking at records or styles, then we'd all be getting rich by those kinds of analysis -- but in fact, as you alluded to in an earlier post, there are so many variables in MMA (and as I stated in earlier posts, in team sports) that the only people who consistently get rich from sports gambling are the bookies and casinos who take in a percentage of the bets.

Be nice if either way worked consistently though -- a few big bets based on record or style and we could retire very young.

Fair enough.

Good talk.