Value of: Bridge Deal RFAs

[Dustin]

Is that true, because all summer I saw Matthews being compared to Crosby and Malkin contracts. The top players in the league will always get paid...I 100% agree.

With these bridge deals, teams are for sure hoping they sign that guy in a few years to a long term deal. The benefit, is that those long term deals end around that age of 30/31...instead of 27/28. So at that point, teams can cut bait with them knowing they got most or all of the prime years from that player. But when faced with signing a 28 year old, like on my team Konecny and Provy, that becomes a lot more questionable on what to do.

But using those 2 as examples, do we really think guys like that are going to be squeezed out and forced to take low contacts? I just don't see it especially when GM's can acquire them for free. The "free" is in the acquiring, the "cost" is often overpaying due to supply and demand.

The negative may be that when they do sign those contracts they far exceed what they could have been had they signed earlier.

Again I believe that context matters in all signings. That being said you have to assume as we get further and further into analytics and ways to evaluate players, GM's will be less inclined to pay players like Neal and Lucic from free agency and more likely to lock up guys like Konecny and Provorov long term.

That being said the nature of the business dictates in a lot of ways how GM's build their teams. If you tell a GM that signing this player now for 8 years may really pay off in 4 years as opposed to if you take a chance on a UFA it may have a positive impact this year, I can see why they would look to the quick fix.

 

[TomasHertlsRooster]

Which means it really doesn't do anything and it's useless, especially in an off-season where players are signing for uncommon terms; something that is not properly adjusted for in the model. Anything you do based on these numbers is equally worthless.

This would only apply to the probability of term number that they calculate. Hypothetically, if they took salary cap situations into account, they would be more successful in predicting what term that a player will take but that shouldn't have any impact on what a player's market value is at that term. And as much as you like to keep saying it, it doesn't make the numbers worthless at all. And every time I've dug deeper into the contracts that players have signed, they seem pretty in line with the historical RFA market outside of few.

 

[Dekes For Days]

Hypothetically, if they took salary cap situations into account, they would be more successful in predicting what term that a player will take but that shouldn't have any impact on what a player's market value is at that term.

Except we can clearly see that term is not accounted for properly in the model. How term affects value is wildly inaccurate, which means it is especially bad to use it in an off-season where uncommon terms are being signed to draw conclusions.

And as much as you like to keep saying it, it doesn't make the numbers worthless at all.

As much as you like denying it because it's the last thing you have to cling to, the numbers are worthless.

 

[TomasHertlsRooster]

Except we can clearly see that term is not accounted for properly in the model. How term affects value is wildly inaccurate, which means it is especially bad to use it in an off-season where uncommon terms are being signed to draw conclusions.


As much as you like denying it because it's the last thing you have to cling to, the numbers are worthless.

The way that term affects valuable is not wildly inaccurate. And this is not the only piece of evidence I have.

You know what's funny here is that you literally said very recently that there has not been a significant shift in the RFA market. So what are you even arguing against?

 

[Dekes For Days]

The way that term affects valuable is not wildly inaccurate.

Lmao. I encourage anybody who wants a good laugh to read this statement by Joe and then look at the values they place at different terms for pretty much any contract.

So what are you even arguing against?

Your use of a horrible projection model to make conclusions you have no business making. And your blatant double standard regarding how this model represents your GM's contracts and other contracts you have spent months ridiculing and mocking.

 

[nbwingsfan]

The teams have zero protection. It's just pushing the problem down the road and hoping that the player won't want to leave/get paid.

The off-season 3 years from now is going to be interesting.

How is it zero protection when they’re still RFA at the end?

 

[TomasHertlsRooster]

Lmao. I encourage anybody who wants a good laugh to read this statement by Joe and then look at the values they place at different terms for pretty much any contract.


Your use of a horrible projection model to make conclusions you have no business making. And your blatant double standard regarding how this model represents your GM's contracts and other contracts you have spent months ridiculing and mocking.

Hmm, let's compare Evolving Hockey's contract projection methods to yours.

You use two cherry picked metrics (ELC ESP/GP and ELC PPP/60) that your favorite players perform well in, and claim that these are the driving factors behind contract valuation. You have NEVER provided any evidence to support this. Meanwhile, I have done your work for you, looked at 5-year post-ELC contract with a cap hit of at least 7%, and found that in a multi-linear regression that accounts for something like ESP/GP, there is no correlation between these factors and pay.



Despite this information, you still continue to tell people that these are the factors that drive contracts in the NHL, and that anybody else who disagrees is wrong. Now, you are ridiculing the contract projections made by Evolving Hockey, despite you providing absolutely ZERO evidence to support your view point that these two metrics are the driving factors behind contracts.

By comparison, the guys at Evolving Hockey created a very intensive contract projection model, which they made an intensive article describing. They created their model with some assistance from Matt Cane, whose 2018 contract projection model had an R^2 in the mid 0.8s. Now, this year's contract projection model from Evolving Hockey has an R^2 of 0.89 with actual contract signing values. For you to ridicule this model in favor of your own "analysis" which focuses on two cherry picked metrics that probably do not even significantly drive pay on their own is HILARIOUS.

You want to try to project contracts for next year's RFA and UFA class, go ahead. I'm sure you will do better than the guys at Evolving Hockey.

 

[Dekes For Days]

Hmm, let's compare Evolving Hockey's contract projection methods to yours.

Which you didn't even do.

You took two of the stats I use, created one all on your own, ignored all context of the contracts despite my constant reminder that context is important, claimed that that is my entire analysis, and then looked at how they compared individually to a tiny sample of cherry-picked contracts that all had very different situations.

Despite this information, you still continue to tell people that these are the factors that drive contracts in the NHL, and that anybody else who disagrees is wrong.

No, actually, the one going around everywhere claiming that they know exactly how contracts are done and anybody that disagrees is wrong is you, as we can all clearly see.

Now, you are ridiculing the contract projections made by Evolving Hockey

No, I'm putting a spotlight on your incorrect and selective use of them to make claims that advance your narrative, while you ignore that the projections have major issues that I'm sure even the creators would admit.

 

[TomasHertlsRooster]

Which you didn't even do.

You took two of the stats I use, created one all on your own, ignored all context of the contracts despite my constant reminder that context is important, claimed that that is my entire analysis, and then looked at how they compared individually to a tiny sample of cherry-picked contracts that all had very different situations.


No, actually, the one going around everywhere claiming that they know exactly how contracts are done and anybody that disagrees is wrong is you, as we can all clearly see.


No, I'm putting a spotlight on your incorrect and selective use of them to make claims that advance your narrative, while you ignore that the projections have major issues that I'm sure even the creators would admit.

Contract year and ELC P/GP combined is the metric that I have been using for months, and it is the one that has the highest correlation with cap hit percentage for RFAs from what I have seen. It does better for predicting contracts than any other metric, and that is why I use it. Meanwhile, after adjusting for this information, metrics like ELC ESP/GP and ELC PPP/60 are close to useless in projecting contracts, yet you continue to act like they are gospel. DESPITE PROVIDING NO EVIDENCE OF THIS.

You call everything cherry picked, but this is pretty much exactly what you asked for. You just don’t like the results so you throw out that word to discredit literally any evidence that somebody else puts out, yet you put out no evidence of your own. Make a real f***ing argument. Use some real f***ing evidence.

The projections have their issues, but on the macro level, their correlation with actual cap hit is an R^2 of 0.89. That shows that they did a pretty damn good job of projecting contracts. And if they did that well, then I think it’s fair to use their model as one piece of evidence to determine whether or not the market has shifted over from historical markets in order to pay RFAs significantly more. Based on the results of every single contract signed, it appears that the market has not shifted at all.

I’m not responding to you until you provide SOME KIND OF EVIDENCE that ELC ESP/GP and ELC PPP/60 are driving factors of pay, more so than something as simple as raw points per game in ELC+ contract year.

 

[Dekes For Days]

Contract year and ELC P/GP combined is the metric that I have been using for months

Yet you claim this is my analysis. Dishonest as usual.

Meanwhile, after adjusting for this information, metrics like ELC ESP/GP and ELC PPP/60 are close to useless in projecting contracts

Gee, it's almost like looking at part of a set of statistics individually with no context in comparison to contracts that rely on multiple pieces of statistics and context will provide inaccurate results. What a surprise.

You've literally proved nothing except that you'll do things wrong to get your desired result. You're the one that keeps bringing up statistics that I use to distract away from your horrible application/use of pretty horrible projection models.

You call everything cherry picked, but this is pretty much exactly what you asked for.

No, it's not at all, and you fully know that.

The projections have their issues

Many major issues that make them worthless for making the conclusions you have.

their correlation with actual cap hit

Correlation =/= causation, yet literally all you do is manipulate numbers and samples until you get desired correlations. And if you still can't make your GM's contracts look good, you just call them special outliers that don't count.

 

[Beyonder91]

Yep as I thought the model was off on Rantanen for 6 year term. Would have been off even more for 8 year I'm sure.
RFA contract are changing.