FYI until last year I used to work in the Local Brand Management industry. Our company had direct relationships with both Google and Yelp.
I can tell you that in years past, we considered Yelp to have higher quality reviews overall. Within the last 3 years or so however, that has begun to shift in Google's favor.
Not in terms of average review length; Yelp still wins there. But in terms of how "reliable" we considered ratings to be for our customers, Google has definitely overtaken Yelp as "The Standard".
Speaking personally as someone who was in that industry for 9 years: I check Google and take it with a grain of salt, but I NEVER bother checking Yelp. Good call-out for Eater though, that's my go-to for cities they've covered.
I do also have some insider knowledge about how Yelp worked with businesses... which doesn't improve trust. For all Google's faults, they just don't care about anyone no matter what, so in that sense they're very egalitarian 😂
This matches my experience as a, uh, prolific reviewer. I used to review places on Yelp, and now I do on Google instead. I think I write quality reviews, but I’m the type of person who edits Wikipedia and includes photos and is thoughtful about what I put in a review.
Google has done a good job gamifying their interface to encourage more info, photos, etc. as well.
Why does QBERT give no credit to YAC? The ability to hit receivers in stride so they can keep moving seems to be a positive skill. I get that a LOT of this is playcalling and receiver skill, but presumably not all of it. Just too hard to extract from the noise?
I am always surprised that the committee doesn't seem to treat an OT win as lesser than a win in regulation. An OT win is more or less a tie where they found a more creative way to flip a coin.
I think the "restaurant recommendations will eventually rely on AI" is strange to me.
Not that I think it's wrong (it feels like an obvious use case for it), but in that it ... inherently cannot try the food or the restaurants.
Not trying to make a stupid remark here, but doesn't this just mean that we're more at the whims of things like Google Reviews / Yelp / critics lists, just with less insight about the who behind the recommendation? Let's say that Yelp stops providing reviews because it goes out of business and critics lists just don't get enough traffic in an AI world to sustain itself, so we're basically just basing this on Gemini surfacing Google Reviews in a personalized way to you. Does Google have to pay for taste-makers to write reviews at that point, and we're just in a similar spot?
Yeah this feels like yet another common good that will be ruined by GenAI. LLMs might be good at collating existing reviews written by humans, but if they take over the space how many humans will continue writing thoughtful reviews for an audience of a couple corporate AI scrapers?
On top of that, the opacity of LLMs will make it extremely tempting to skew the results in exchange for payment and I fully expect the big tech companies producing these bots to do exactly that once they've taken over the market; much like they have done with all of the other markets they dominate.
SBSQ/article ??- Would love an article on Michigan's Gubernatorial race this year.
You have Whitmer term limited in a purple state where the SoS & Attorney General are also Dems & term limited. & an open Senate seat from Dem. Gary Peters retiring.
Whitmer's Lt. Gov has already dropped out of the race, but, SoS Jocelyn Benson is the likely Dem nominee. Not shockingly, a # of major GOP people are running, too.
What makes it interesting, Detroit Mayor Mike Duggan, who is a Democrat, is running as an independent & as of today, leads the way-too-early polls. Lots of sniping between him & other Dems, which tells me they're concerned his presence will allow the GOP to win the election, even in a tough year.
I'll admit I'm a Duggan guy- I live near Detroit & the job he's done has been phenomenal. He's a Dem, but, not sure that wasn't just due to being a requisite to winning in Detroit. He has governed like an old school moderate Dem or Republican, which is a breath of fresh air in these partisan times.
Anyway, would love thoughts/predictions on the race.
SBSQ #29 - With the increasing ease in betting on sports (including prediction markets) will we see increasing fixing in some events when a single player can make a difference? We have just seen another major basketball betting scandal which looks to be international in nature. I'm skeptical that anything can be done about this, what is your take?
In "Kamala Harris has Liz Cheney Syndrome", you lay out the ideological headwinds Harris faced in 2024. One thing that's often puzzled me was the suggestion that Biden was unfit to campaign, but somehow still fit to serve. Had he resigned, would President Harris have had a better shot at winning election that Vice President Harris?
I agree the CFP committee needs to have an explicit objective function, but it needn’t be one dimensional (I don’t think their implicit objectives are). Without an explicit objective function, it’s likely that different committees members are using different objective functions, which may account for some of the inconsistency. The objective function is likely something like x*deservedness + (1-x) *probability of winning, possibly with minimums for each metric. Of course, then they need to have a clear criteria for deservedness—something like how other teams would have likely performed against a team’s schedule. It is also reasonable, if made explicit, to treat conference championship games differently to avoid punishing teams for making the game. The basic problem is lack of rigor in setting the objectives. There still would be disagreements about how the terms are evaluated, but at least everyone would know what the argument was about.
Re: strength of schedule for Stafford v. Maye, it’d be nice to have a breakdown of the EPA vs. DVOA debate (former seems to favor Maye and latter Stafford?)
Very old school of me, but I miss the printed version of Zagat's, which we relied on for years. Now I live in Maine in essentially a restaurant desert, so we settle for anything decent and mostly cook at home!
Matt tweeted that he's been sick for several days. [https://x.com/mattyglesias/status/2013373445115941251] Proof of how sick he feels is that he's hardly even been tweeting. Hope he feels well enough by tomorrow to have the energy to be able to chat with Nate.
I don't even see a reasonable debate: the college football playoffs, like all sports playoffs, should be based on desserts, not prediction. We have games to decide winners, not to gather data for estimating which team would win more games in 100 replays. Win by luck, win by skill, you're still the champion. Imagine someone suggesting that we ignore the regular season in baseball and let Las Vegas oddsmakers choose who will play in the World Series.
Similarly, I don't understand the complaint that there aren't enough games to establish the best team reliably. That's not the point of sport. In any case, there probably isn't a consensus best team. Different teams are better at different points in the season, or under different game conditions, or versus different types of opponents. One team might be most reliable in beating average teams, another the favorite in head-to-head matches against the other top teams.
In terms of what the selection committee actually does, I'd say it ignores both desserts and prediction to deliver the slate that maximizes revenue. It's more like WWE than a sports championship--or more precisely, like the rest of the bowl games. Of course that's what you'd expect them to do, at least if you're a University of Chicago guy like me.
The biggest single flaw I see is game results are evaluated in light of strength of opponent using in-season results. A desserts system would consider strength of opponent evaluated when the games were scheduled. It's not fair to penalize a team that put hard games on its schedule if some of those opponents were weaker than expected, or to reward a team that chose easy opponents that turned out to be surprisingly tough.
In the past we used to say the barrier to more rational scheduling that would make the most deserving (also the best) teams more obvious was traditional rivalries and conference commitments. Athletic departments claimed to want to tie football game opponents to other sports as well. But college football has smashed through those kinds of considerations since CoVID and we could move to better scheduling for the teams that wanted to compete for the national championship. The Ivies, and military schools and other less competitive conferences could continue as they were.
FYI until last year I used to work in the Local Brand Management industry. Our company had direct relationships with both Google and Yelp.
I can tell you that in years past, we considered Yelp to have higher quality reviews overall. Within the last 3 years or so however, that has begun to shift in Google's favor.
Not in terms of average review length; Yelp still wins there. But in terms of how "reliable" we considered ratings to be for our customers, Google has definitely overtaken Yelp as "The Standard".
Speaking personally as someone who was in that industry for 9 years: I check Google and take it with a grain of salt, but I NEVER bother checking Yelp. Good call-out for Eater though, that's my go-to for cities they've covered.
I do also have some insider knowledge about how Yelp worked with businesses... which doesn't improve trust. For all Google's faults, they just don't care about anyone no matter what, so in that sense they're very egalitarian 😂
This matches my experience as a, uh, prolific reviewer. I used to review places on Yelp, and now I do on Google instead. I think I write quality reviews, but I’m the type of person who edits Wikipedia and includes photos and is thoughtful about what I put in a review.
Google has done a good job gamifying their interface to encourage more info, photos, etc. as well.
For next SBSQ: can you give the behind the scenes story of Risky Business? And what led to its end? What’s next for you in podcasting?
Came here to ask this. I hope Nate and Maria keep putting out poker content together.
Strongly seconded.
I would also love to see new Nate podcast with different co-host(s). Either some of the old 538 crew or someone new.
All tacos are better in San Diego, not just fish
Why does QBERT give no credit to YAC? The ability to hit receivers in stride so they can keep moving seems to be a positive skill. I get that a LOT of this is playcalling and receiver skill, but presumably not all of it. Just too hard to extract from the noise?
You are going to be completely wrong when my Broncos led by Stiddy beat Maye
I am always surprised that the committee doesn't seem to treat an OT win as lesser than a win in regulation. An OT win is more or less a tie where they found a more creative way to flip a coin.
I think the "restaurant recommendations will eventually rely on AI" is strange to me.
Not that I think it's wrong (it feels like an obvious use case for it), but in that it ... inherently cannot try the food or the restaurants.
Not trying to make a stupid remark here, but doesn't this just mean that we're more at the whims of things like Google Reviews / Yelp / critics lists, just with less insight about the who behind the recommendation? Let's say that Yelp stops providing reviews because it goes out of business and critics lists just don't get enough traffic in an AI world to sustain itself, so we're basically just basing this on Gemini surfacing Google Reviews in a personalized way to you. Does Google have to pay for taste-makers to write reviews at that point, and we're just in a similar spot?
Yeah this feels like yet another common good that will be ruined by GenAI. LLMs might be good at collating existing reviews written by humans, but if they take over the space how many humans will continue writing thoughtful reviews for an audience of a couple corporate AI scrapers?
On top of that, the opacity of LLMs will make it extremely tempting to skew the results in exchange for payment and I fully expect the big tech companies producing these bots to do exactly that once they've taken over the market; much like they have done with all of the other markets they dominate.
Nate - why does QBERT hate Caleb Williams so much when he’s obviously awesome?
SBSQ/article ??- Would love an article on Michigan's Gubernatorial race this year.
You have Whitmer term limited in a purple state where the SoS & Attorney General are also Dems & term limited. & an open Senate seat from Dem. Gary Peters retiring.
Whitmer's Lt. Gov has already dropped out of the race, but, SoS Jocelyn Benson is the likely Dem nominee. Not shockingly, a # of major GOP people are running, too.
What makes it interesting, Detroit Mayor Mike Duggan, who is a Democrat, is running as an independent & as of today, leads the way-too-early polls. Lots of sniping between him & other Dems, which tells me they're concerned his presence will allow the GOP to win the election, even in a tough year.
I'll admit I'm a Duggan guy- I live near Detroit & the job he's done has been phenomenal. He's a Dem, but, not sure that wasn't just due to being a requisite to winning in Detroit. He has governed like an old school moderate Dem or Republican, which is a breath of fresh air in these partisan times.
Anyway, would love thoughts/predictions on the race.
SBSQ #29 - With the increasing ease in betting on sports (including prediction markets) will we see increasing fixing in some events when a single player can make a difference? We have just seen another major basketball betting scandal which looks to be international in nature. I'm skeptical that anything can be done about this, what is your take?
In "Kamala Harris has Liz Cheney Syndrome", you lay out the ideological headwinds Harris faced in 2024. One thing that's often puzzled me was the suggestion that Biden was unfit to campaign, but somehow still fit to serve. Had he resigned, would President Harris have had a better shot at winning election that Vice President Harris?
I agree the CFP committee needs to have an explicit objective function, but it needn’t be one dimensional (I don’t think their implicit objectives are). Without an explicit objective function, it’s likely that different committees members are using different objective functions, which may account for some of the inconsistency. The objective function is likely something like x*deservedness + (1-x) *probability of winning, possibly with minimums for each metric. Of course, then they need to have a clear criteria for deservedness—something like how other teams would have likely performed against a team’s schedule. It is also reasonable, if made explicit, to treat conference championship games differently to avoid punishing teams for making the game. The basic problem is lack of rigor in setting the objectives. There still would be disagreements about how the terms are evaluated, but at least everyone would know what the argument was about.
Re: strength of schedule for Stafford v. Maye, it’d be nice to have a breakdown of the EPA vs. DVOA debate (former seems to favor Maye and latter Stafford?)
Very old school of me, but I miss the printed version of Zagat's, which we relied on for years. Now I live in Maine in essentially a restaurant desert, so we settle for anything decent and mostly cook at home!
Matt tweeted that he's been sick for several days. [https://x.com/mattyglesias/status/2013373445115941251] Proof of how sick he feels is that he's hardly even been tweeting. Hope he feels well enough by tomorrow to have the energy to be able to chat with Nate.
I don't even see a reasonable debate: the college football playoffs, like all sports playoffs, should be based on desserts, not prediction. We have games to decide winners, not to gather data for estimating which team would win more games in 100 replays. Win by luck, win by skill, you're still the champion. Imagine someone suggesting that we ignore the regular season in baseball and let Las Vegas oddsmakers choose who will play in the World Series.
Similarly, I don't understand the complaint that there aren't enough games to establish the best team reliably. That's not the point of sport. In any case, there probably isn't a consensus best team. Different teams are better at different points in the season, or under different game conditions, or versus different types of opponents. One team might be most reliable in beating average teams, another the favorite in head-to-head matches against the other top teams.
In terms of what the selection committee actually does, I'd say it ignores both desserts and prediction to deliver the slate that maximizes revenue. It's more like WWE than a sports championship--or more precisely, like the rest of the bowl games. Of course that's what you'd expect them to do, at least if you're a University of Chicago guy like me.
The biggest single flaw I see is game results are evaluated in light of strength of opponent using in-season results. A desserts system would consider strength of opponent evaluated when the games were scheduled. It's not fair to penalize a team that put hard games on its schedule if some of those opponents were weaker than expected, or to reward a team that chose easy opponents that turned out to be surprisingly tough.
In the past we used to say the barrier to more rational scheduling that would make the most deserving (also the best) teams more obvious was traditional rivalries and conference commitments. Athletic departments claimed to want to tie football game opponents to other sports as well. But college football has smashed through those kinds of considerations since CoVID and we could move to better scheduling for the teams that wanted to compete for the national championship. The Ivies, and military schools and other less competitive conferences could continue as they were.