Auction Value Strategy (1 Viewer)

[azmat]

The following auction value methodology has been used for the past decade in my home league and wanted to share in case it provides value to others. Examples for my league are in parenthesis:

Using previous year's data, take the first ADP for each position that goes for $1 in the draft. (QB21, RB52, WR55, TE18).

Using previous year's data, list the points per game for each position by ADP and the percentage that each ppg is relative to the sum of the overall position.

Calculate the VORP for each player by subtracting the respective ppg from the ADP from the first player in that position that goes for $1. (QB1 = 30.3 - 16.1 = 14.2 VORP)

Add up the VORPs for each of the positions (QB=102, RB=288, WR=287, TE=51).

Calculate the VORP percentage by position (QB=102/729=14%)

Multiply total auction dollars by this percentage for each position.

For each position, multiply this number by number of players you plan to roster for each starting position.

Sum up the four numbers for QB, RB, WR, TE

Take this number and divide by your total auction dollars.

Take the total VORP for each position and divide by number from step 9. This is the number of auction dollars that should be allocated to each position by the league.

Finally, take the percentage of each player from step 2 and multiply the respective position value from step 10. This is the value that each player should be assigned.

I like this method, because it leverages historical data, with ADP and quantity of starting positions. Happy to share more details if there is interest or answer any questions.

 

[BroncoFreak_2K3]

The following auction value methodology has been used for the past decade in my home league and wanted to share in case it provides value to others. Examples for my league are in parenthesis:

Using previous year's data, take the first ADP for each position that goes for $1 in the draft. (QB21, RB52, WR55, TE18).

Using previous year's data, list the points per game for each position by ADP and the percentage that each ppg is relative to the sum of the overall position.

Calculate the VORP for each player by subtracting the respective ppg from the ADP from the first player in that position that goes for $1. (QB1 = 30.3 - 16.1 = 14.2 VORP)

Add up the VORPs for each of the positions (QB=102, RB=288, WR=287, TE=51).

Calculate the VORP percentage by position (QB=102/729=14%)

Multiply total auction dollars by this percentage for each position.

For each position, multiply this number by number of players you plan to roster for each starting position.

Sum up the four numbers for QB, RB, WR, TE

Take this number and divide by your total auction dollars.

Take the total VORP for each position and divide by number from step 9. This is the number of auction dollars that should be allocated to each position by the league.

Finally, take the percentage of each player from step 2 and multiply the respective position value from step 10. This is the value that each player should be assigned.

I like this method, because it leverages historical data, with ADP and quantity of starting positions. Happy to share more details if there is interest or answer any questions.

Awesome stuff. Between this and Scoresmansw VORP thread, this is the best auction strategy information I’ve seen on this site, including staff articles. Sorry Joe


More details if you have them, yes please

 

[azmat]

I'd be happy to share a spreadsheet for my league if that is helpful and thanks for the feedback.

 

[Scoresman]

Yeah I do something similar. This is valuable stuff.

I also compile my leagues historical auction prices and use those to come up with estimates. I look at these vs the VORP “budgets” and can usually easily see where value will most likely appear during the auction. It’s a big help to identify these before the auction even begins.

 

[BroncoFreak_2K3]

I'd be happy to share a spreadsheet for my league if that is helpful and thanks for the feedback.

Yes please. Thanks for the great info!

I try to do something similar every year, but having a nice template that uses a good, sensible process to start with like this would be very a valuable tool.

 

[Payne]

This is awesome. I have found the VBD excel tool on FBG to be very accurate in determining auction values.
What I do is very simple. I set my rankings and then use a moving average of what the top QB's, RB's, etc went for over the past several years and apply them to my rankings. It's quick, simple and has been fairly spot on.

I will try your method.

 

[bro1ncos]

I'd be happy to share a spreadsheet for my league if that is helpful and thanks for the feedback.

I would appreciate seeing this

 

[azmat]

The 'vorp.xls' file has been uploaded to the link here: https://easyupload.io/euuino
Feel free to reach out with any questions.

 

[Gally]

Using previous year's data, take the first ADP for each position that goes for $1 in the draft. (QB21, RB52, WR55, TE18).

So are you saying that the QB21 last year (based on final actual point totals) went for $1 as did QB22, 23, etc? QB 20 (based on final point totals) went for more than $1?

 

[azmat]

For my particular league, which is a 14-teamer, 'AV" are historical auction values using the past 8 years of my league's draft data.

 

[Gally]

For my particular league, which is a 14-teamer, 'AV" are historical auction values using the past 8 years of my league's draft data.

But is that based on the actual finish of the QB or the QB that was takes as QB21 in the auction (based on price of $1)?

The reason I am asking is because in the "redraft mistake" thread we had a discussion about ADP and it got mixed in with final finish. I just want to be clear if this is using finishes for QB21 or the actual 21st QB taken (by auction price).

(not sure if this makes sense and I apologize if it doesn't)

 

[azmat]

Makes senses. AV is the average auction value of the 21st selected QB in the draft (pre-season) for this league.

 

[BroncoFreak_2K3]

How would you compensate for keepers with this process? They are kinda outside of the auction process somewhat as we load those in pre-draft. However, their keeper values are subtracted from each team's total auction budgets.

For my auction league, keepers are kept for their previous year's auction value * 25% inflation annually and can be kept by the same team up to 2 consecutive years.

 

[azmat]

I have never looked at keepers, but I will take a look this week to see if the formulas can be adjusted in some manner to compensate.

 

[Scoresman]

My method for coming up with the "budgets" for each player is similar but I calculate a bit different. I use pretty standard VBD system. I choose my baselines based on a google sheet someone on reddit made that uses historical points and games.

Man-Games VBD Baselines (2023)

docs.google.com

I add up all of the players' VBD values.
I then calculate my "extra dollars" which is every dollar available to spend after paying $1 for each roster spot.
I divide these two to get Pts/Extra Dollar

I then divide every player's VBD by the Pts/Extra Dollar ratio (and add in the $1 minimum) to get the budgets. Nice check here is that your pool of players' prices should all add up to your total dollars to be spent.

 

[CalBear]

The following auction value methodology has been used for the past decade in my home league and wanted to share in case it provides value to others. Examples for my league are in parenthesis:

This is pretty similar to the setup I use, with one substantial difference. In my view, linear assignment of VORP points overvalues mediocre players and undervalues studs, partly because of the cost of roster spots, partly because of the higher reliability of projections for top players. My general rule of thumb in auctions is, if you're bidding on the best player left on the board, win the auction. Paying $24 for a $20 player is better than getting two $12 players at par value. I wanted my calculations to reflect that strategy, so I set up a sigmoid curve function.

The calculation is:

• VBDS = (VBD - (crossover)) / (sigmoid denominator) [In my league I start with 80, 20 but I tweak them]
• EXPZ = 1 / (1 + EXP (-VBDS)) [Sigmoid function based on VBDS. Equals zero when VBD = crossover].
• ValueS = (bid $ available) / (sum of EXPZ) * EXPZ

The crossover parameter sets the point at which the sigmoid curve crosses over from increasing to decreasing slope; I set it to a VBD number which represents, "after this, players get less interesting." In 2023 that was 50% of the top VBD number, which was guys like Rhamondre Stevenson and Jaylen Waddle. The calculation comes out to have those at about $10 of a $100 budget.

The sigmoid denominator sets the steepness of the curve. With it set at 20, the top projected player (which was Ekeler in our PPR league last year came out at $39. Set it to 15 and he's worth $41; set it to 25 and he's worth $37. I play around until I like the shape of the curve.

One thing to note when doing these calculations: The total bid $ available for a given player isn't "# of teams * budget". Assuming everyone costs at least $1, it's "# of teams * budget - number of player slots", because all teams have to keep enough money to fill their rosters.

 

[BroncoFreak_2K3]

This is all really interesting great stuff. I certainly appreciate you guys sharing your strategy setups.

I've been wrestling with an issue that always comes up in my auction league, and I'd love to find a solution, or a way to alleviate it from affecting a good size chuck of our draft, and the time it takes to go through this part of the draft, at the end of the draft.

So for my 10-team auction draft (we draft a total of around 230 total players (which includes Defenses). We carry 25 player rosters, yea, it's a very deep league and large starting rosters (13 total starters). Its a really fun league that has lasted many years and there has been little to no turnover. Everyone loves it.

So for about the first 2/3rds of the draft, all teams tend to spend pretty liberally. Then for about the last third of the draft, we deal with the majority of teams having very little auction $ left. This translates to anywhere from 40-60 picks being $1-$2 winning bids. The positive for it is that almost all teams are able to find some value in players selected during this time of the draft and they turn out to be keepers the following season (Last 2 seasons players like: Trevor Lawrence, Joe Burrow, Jared Goff, George Pickens, Raheem Mostert, Bryce Young, Rashee Rice, Jaylen Warren)

The downside always is waiting for the bid clock to expire (20 seconds, resets every time bid is increased).

I'd love to figure out a way to eliminate this dead period at the end of our drafts every year. Shortening rosters and benches will still translate to heavy spending for the first 2/3rds of the draft I'm sure, so I'm looking for other alternatives. I really don't like having this much of our draft taken up with $1 and $2 players.

Any ideas out there? Appreciate the time and this great thread. Thanks

 

[dkocourek]

This is all really interesting great stuff. I certainly appreciate you guys sharing your strategy setups.

I've been wrestling with an issue that always comes up in my auction league, and I'd love to find a solution, or a way to alleviate it from affecting a good size chuck of our draft, and the time it takes to go through this part of the draft, at the end of the draft.

So for my 10-team auction draft (we draft a total of around 230 total players (which includes Defenses). We carry 25 player rosters, yea, it's a very deep league and large starting rosters (13 total starters). Its a really fun league that has lasted many years and there has been little to no turnover. Everyone loves it.

So for about the first 2/3rds of the draft, all teams tend to spend pretty liberally. Then for about the last third of the draft, we deal with the majority of teams having very little auction $ left. This translates to anywhere from 40-60 picks being $1-$2 winning bids. The positive for it is that almost all teams are able to find some value in players selected during this time of the draft and they turn out to be keepers the following season. The downside always is waiting for the bid clock to expire (20 seconds, resets every time bid is increased).

I'd love to figure out a way to eliminate this dead period at the end of our drafts every year. Shortening rosters and benches will still translate to heavy spending for the first 2/3rds of the draft I'm sure, so I'm looking for other alternatives. I really don't like having this much of our draft taken up with $1 and $2 players.

Any ideas out there? Appreciate the time and this great thread. Thanks

This has been our league’s issue too. Our rosters are 17 deep. We decided to do a $100 auction for 10 players each, with a snake draft for the remaining 7 players.

 

[CalBear]

This is all really interesting great stuff. I certainly appreciate you guys sharing your strategy setups.

I've been wrestling with an issue that always comes up in my auction league, and I'd love to find a solution, or a way to alleviate it from affecting a good size chuck of our draft, and the time it takes to go through this part of the draft, at the end of the draft.

So for my 10-team auction draft (we draft a total of around 230 total players (which includes Defenses). We carry 25 player rosters, yea, it's a very deep league and large starting rosters (13 total starters). Its a really fun league that has lasted many years and there has been little to no turnover. Everyone loves it.

So for about the first 2/3rds of the draft, all teams tend to spend pretty liberally. Then for about the last third of the draft, we deal with the majority of teams having very little auction $ left. This translates to anywhere from 40-60 picks being $1-$2 winning bids. The positive for it is that almost all teams are able to find some value in players selected during this time of the draft and they turn out to be keepers the following season (Last 2 seasons players like: Trevor Lawrence, Joe Burrow, Jared Goff, George Pickens, Raheem Mostert, Bryce Young, Rashee Rice, Jaylen Warren)

The downside always is waiting for the bid clock to expire (20 seconds, resets every time bid is increased).

I'd love to figure out a way to eliminate this dead period at the end of our drafts every year. Shortening rosters and benches will still translate to heavy spending for the first 2/3rds of the draft I'm sure, so I'm looking for other alternatives. I really don't like having this much of our draft taken up with $1 and $2 players.

Any ideas out there? Appreciate the time and this great thread. Thanks

I think the combined auction/snake is a reasonable way to approach it. The reality is, half of the players are not worth more than $2. If you have a 25-player roster that starts 6 skill position players, the team that spends $80 on its top 6 players will very likely beat the teams that spend $60 on them.

You could change that dynamic a little bit by expanding the starting positions, a la Anarchy: 25 players but you start 13 of them.

 

[BroncoFreak_2K3]

This is all really interesting great stuff. I certainly appreciate you guys sharing your strategy setups.

I've been wrestling with an issue that always comes up in my auction league, and I'd love to find a solution, or a way to alleviate it from affecting a good size chuck of our draft, and the time it takes to go through this part of the draft, at the end of the draft.

So for my 10-team auction draft (we draft a total of around 230 total players (which includes Defenses). We carry 25 player rosters, yea, it's a very deep league and large starting rosters (13 total starters). Its a really fun league that has lasted many years and there has been little to no turnover. Everyone loves it.

So for about the first 2/3rds of the draft, all teams tend to spend pretty liberally. Then for about the last third of the draft, we deal with the majority of teams having very little auction $ left. This translates to anywhere from 40-60 picks being $1-$2 winning bids. The positive for it is that almost all teams are able to find some value in players selected during this time of the draft and they turn out to be keepers the following season (Last 2 seasons players like: Trevor Lawrence, Joe Burrow, Jared Goff, George Pickens, Raheem Mostert, Bryce Young, Rashee Rice, Jaylen Warren)

The downside always is waiting for the bid clock to expire (20 seconds, resets every time bid is increased).

I'd love to figure out a way to eliminate this dead period at the end of our drafts every year. Shortening rosters and benches will still translate to heavy spending for the first 2/3rds of the draft I'm sure, so I'm looking for other alternatives. I really don't like having this much of our draft taken up with $1 and $2 players.

Any ideas out there? Appreciate the time and this great thread. Thanks

I think the combined auction/snake is a reasonable way to approach it. The reality is, half of the players are not worth more than $2. If you have a 25-player roster that starts 6 skill position players, the team that spends $80 on its top 6 players will very likely beat the teams that spend $60 on them.

You could change that dynamic a little bit by expanding the starting positions, a la Anarchy: 25 players but you start 13 of them.

Thanks. The combo draft idea could work.

We already have pretty sizable starting rosters: 2QB, 2RB, 4WR, 2TE, 1FLEX ( RB, WR, TE), 1 K, 1DEF.

 

[Instinctive]

I'll have to go into my old spreadsheet, but I did something slightly different, which is to try and incorporate an outcome distribution in projections and also raise that VORP baseline to be the 25th or 50th percentile starter instead of last starter. Because I want a dominant team.

 

[azmat]

This is a very interesting approach and I have never thought of raising the VORP baseline higher than last starter. I will take a look at this over the next couple of days. Thanks for sharing!

 

[azmat]

The following auction value methodology has been used for the past decade in my home league and wanted to share in case it provides value to others. Examples for my league are in parenthesis:

This is pretty similar to the setup I use, with one substantial difference. In my view, linear assignment of VORP points overvalues mediocre players and undervalues studs, partly because of the cost of roster spots, partly because of the higher reliability of projections for top players. My general rule of thumb in auctions is, if you're bidding on the best player left on the board, win the auction. Paying $24 for a $20 player is better than getting two $12 players at par value. I wanted my calculations to reflect that strategy, so I set up a sigmoid curve function.

The calculation is:

• VBDS = (VBD - (crossover)) / (sigmoid denominator) [In my league I start with 80, 20 but I tweak them]
• EXPZ = 1 / (1 + EXP (-VBDS)) [Sigmoid function based on VBDS. Equals zero when VBD = crossover].
• ValueS = (bid $ available) / (sum of EXPZ) * EXPZ

The crossover parameter sets the point at which the sigmoid curve crosses over from increasing to decreasing slope; I set it to a VBD number which represents, "after this, players get less interesting." In 2023 that was 50% of the top VBD number, which was guys like Rhamondre Stevenson and Jaylen Waddle. The calculation comes out to have those at about $10 of a $100 budget.

The sigmoid denominator sets the steepness of the curve. With it set at 20, the top projected player (which was Ekeler in our PPR league last year came out at $39. Set it to 15 and he's worth $41; set it to 25 and he's worth $37. I play around until I like the shape of the curve.

One thing to note when doing these calculations: The total bid $ available for a given player isn't "# of teams * budget". Assuming everyone costs at least $1, it's "# of teams * budget - number of player slots", because all teams have to keep enough money to fill their rosters.

Is the sigmoid denominator a percentage (0.2) or whole number (20)?

 

[Mr.Pack]

The following auction value methodology has been used for the past decade in my home league and wanted to share in case it provides value to others. Examples for my league are in parenthesis:

Using previous year's data, take the first ADP for each position that goes for $1 in the draft. (QB21, RB52, WR55, TE18).

Using previous year's data, list the points per game for each position by ADP and the percentage that each ppg is relative to the sum of the overall position.

Calculate the VORP for each player by subtracting the respective ppg from the ADP from the first player in that position that goes for $1. (QB1 = 30.3 - 16.1 = 14.2 VORP)

Add up the VORPs for each of the positions (QB=102, RB=288, WR=287, TE=51).

Calculate the VORP percentage by position (QB=102/729=14%)

Multiply total auction dollars by this percentage for each position.

For each position, multiply this number by number of players you plan to roster for each starting position.

Sum up the four numbers for QB, RB, WR, TE

Take this number and divide by your total auction dollars.

Take the total VORP for each position and divide by number from step 9. This is the number of auction dollars that should be allocated to each position by the league.

Finally, take the percentage of each player from step 2 and multiply the respective position value from step 10. This is the value that each player should be assigned.

I like this method, because it leverages historical data, with ADP and quantity of starting positions. Happy to share more details if there is interest or answer any questions.

Awesome stuff. Between this and Scoresmansw VORP thread, this is the best auction strategy information I’ve seen on this site, including staff articles. Sorry Joe


More details if you have them, yes please

But…… is this proven to give good results? Auctions always go sideways eventually and you have to overpay at all positions.

 

[Mr.Pack]

This is all really interesting great stuff. I certainly appreciate you guys sharing your strategy setups.

I've been wrestling with an issue that always comes up in my auction league, and I'd love to find a solution, or a way to alleviate it from affecting a good size chuck of our draft, and the time it takes to go through this part of the draft, at the end of the draft.

So for my 10-team auction draft (we draft a total of around 230 total players (which includes Defenses). We carry 25 player rosters, yea, it's a very deep league and large starting rosters (13 total starters). Its a really fun league that has lasted many years and there has been little to no turnover. Everyone loves it.

So for about the first 2/3rds of the draft, all teams tend to spend pretty liberally. Then for about the last third of the draft, we deal with the majority of teams having very little auction $ left. This translates to anywhere from 40-60 picks being $1-$2 winning bids. The positive for it is that almost all teams are able to find some value in players selected during this time of the draft and they turn out to be keepers the following season. The downside always is waiting for the bid clock to expire (20 seconds, resets every time bid is increased).

I'd love to figure out a way to eliminate this dead period at the end of our drafts every year. Shortening rosters and benches will still translate to heavy spending for the first 2/3rds of the draft I'm sure, so I'm looking for other alternatives. I really don't like having this much of our draft taken up with $1 and $2 players.

Any ideas out there? Appreciate the time and this great thread. Thanks

This has been our league’s issue too. Our rosters are 17 deep. We decided to do a $100 auction for 10 players each, with a snake draft for the remaining 7 players.

But if it’s an auction league don’t you have a set salary cap for the year? I kind of like this idea but we have a set cap and a default cap number per position. How would you incorporate that, or don’t you have a cap?

 

[Scoresman]

Wondering what the brain trust's thoughts are on this VBD baseline strategy.

It takes the popular "Value Based Drafting Revisited" article from 2012 and updates it with current data.

https://www.reddit.com/r/fantasyfootball/comments/1ejw4wa/2024_valuebased_draft_baselines/

 

[RenegadeRoy]

I'm too dumb to understand most of this stuff, but it does interest me and I think it's more or less what I'm doing with my prep (look at previous years projections then how much players went for). Still the number crunching is very interesting to me.

But like Mr. Pack said, auctions have a way of going sideways so my #1 auction strategy is have a plan of some sort. The most simple one I go with is "spend somewhere around half your budget for your top two guys who are ideally in the top 10-20ish ranked players" (for a 12 person league). Keep a little purse for those "middle round" guys, but don't be too stingy early. Tiers lists are much more valuable in auction to me than in snake, so have those ironed out. Give a rough estimate of how much you want to spend for each tier. And of course the golden rule of "know your league."

 

[CalBear]

I'll have to go into my old spreadsheet, but I did something slightly different, which is to try and incorporate an outcome distribution in projections and also raise that VORP baseline to be the 25th or 50th percentile starter instead of last starter. Because I want a dominant team.

That's how I used to tweak the old Draft Dominator app (not possible in the current version I think); moving the baselines up to set the first player over $1 to be someone decent. The problem is that it felt like that method wound up overvaluing the middle of the curve, contributor-level players who

The following auction value methodology has been used for the past decade in my home league and wanted to share in case it provides value to others. Examples for my league are in parenthesis:

This is pretty similar to the setup I use, with one substantial difference. In my view, linear assignment of VORP points overvalues mediocre players and undervalues studs, partly because of the cost of roster spots, partly because of the higher reliability of projections for top players. My general rule of thumb in auctions is, if you're bidding on the best player left on the board, win the auction. Paying $24 for a $20 player is better than getting two $12 players at par value. I wanted my calculations to reflect that strategy, so I set up a sigmoid curve function.

The calculation is:

• VBDS = (VBD - (crossover)) / (sigmoid denominator) [In my league I start with 80, 20 but I tweak them]
• EXPZ = 1 / (1 + EXP (-VBDS)) [Sigmoid function based on VBDS. Equals zero when VBD = crossover].
• ValueS = (bid $ available) / (sum of EXPZ) * EXPZ

The crossover parameter sets the point at which the sigmoid curve crosses over from increasing to decreasing slope; I set it to a VBD number which represents, "after this, players get less interesting." In 2023 that was 50% of the top VBD number, which was guys like Rhamondre Stevenson and Jaylen Waddle. The calculation comes out to have those at about $10 of a $100 budget.

The sigmoid denominator sets the steepness of the curve. With it set at 20, the top projected player (which was Ekeler in our PPR league last year came out at $39. Set it to 15 and he's worth $41; set it to 25 and he's worth $37. I play around until I like the shape of the curve.

One thing to note when doing these calculations: The total bid $ available for a given player isn't "# of teams * budget". Assuming everyone costs at least $1, it's "# of teams * budget - number of player slots", because all teams have to keep enough money to fill their rosters.

Is the sigmoid denominator a percentage (0.2) or whole number (20)?

Whole number. In the spreadsheet it's =(E2-$C$446)/$C$447: E2 is the VBD projection for the player, C446 is the crossover, C447 is the sigmoid denominator. Basically changes the slope of the curve once you feed the result into the sigmoid function. Last year in my league the VBDS function ranged from +4.02 to -4.00. (For players with VBD<=0, VBDS = ((0 - crossover) / sigmoid denominator), so using 80 and 20 comes out to -4.00).

 

[Ray_T]

in the end, when bidding on a player your original prelim vallues can be somewhat meaningless. sometimes you need to go above your max value to get the player. but it is important to know when to walk away.

sometimes its easier to see where the value is once a couple of players have gone and you have to adjust your prelim values because they are a couple dollars too high (or too low)

usually I have found that if your prelim value for a top WR is off by 2 or 3 bucks, then its likely the same for most WR of comparable value. so you adjust your estimate and try to figure out which players become worth less. (if WR are undervalued, chances are another position is overvalued)

 

[Instinctive]

in the end, when bidding on a player your original prelim vallues can be somewhat meaningless. sometimes you need to go above your max value to get the player. but it is important to know when to walk away.

sometimes its easier to see where the value is once a couple of players have gone and you have to adjust your prelim values because they are a couple dollars too high (or too low)

usually I have found that if your prelim value for a top WR is off by 2 or 3 bucks, then its likely the same for most WR of comparable value. so you adjust your estimate and try to figure out which players become worth less. (if WR are undervalued, chances are another position is overvalued)

This is what a dynamic pricing model is good for. Set it up so that as soon as a player is won, values adjust for the rest based on points/position and such.

 

[Scoresman]

Here's my thought process this year now that I have both VBD-based Auction budgets and player price estimates based on historical league data.

I always start this part mocking a few teams using the price estimates, and then listing out what I want based on the best looking mock teams. This year, it's settling into the following. This is with a salary cap of $110.

1.) Two core stud WRs and the best RB I can get for less than $25. According to my budget vs. estimate, some of these players should come at me with some value.
This should net me a core of something like the following

Tyreek Hill $27
Amon-Ra $20
Derrick Henry $24

I feel like that's a good core and gives me $39 for the rest of my team, which is low, but I can manage. I've done it before.

2.) Next up is QB. Because of the money I spent on my core, I have to spend down at QB. This is where it gets tough. I know I want a running QB. Looking at what I can afford, this leaves Kyler Murray at about $5, or Jayden Daniels at about $3. I really wish there were more cheap, running QB options here. I would prefer Kyler as the known commodity, but I hate having just one player to target to make my team plan work. And if just one other team has the same idea, both of them could go well above budget. I suppose I could settle for one of the cheap pocket passing QBs.

3.) TE. Another problem. All of the experts I follow seems to be suggesting spending up at TE. The cheapest of the stud tier of TE should go for $9. This year, looks like it's Kittle as my TE5. Let's roll with it for now. So the team now looks like

Kyler Murray $5
Tyreek Hill $27
Amon-Ra $20
Derrick Henry $24
Kittle $9

4.) That's a good core, now let's fill in the rest of the starters with good value cheap players

Kyler Murray $5
Derrick Henry $24
David Montgomery $11
Tyreek Hill $27
Amon-Ra $20
Amari Cooper $7
Kittle $9

This leaves $1 for K, DEF, and all bench slots. I'm fine with this because free agency is so much more important when it comes to filling out the bench.

I would be ecstatic with the team above.

 

[azmat]

I'll have to go into my old spreadsheet, but I did something slightly different, which is to try and incorporate an outcome distribution in projections and also raise that VORP baseline to be the 25th or 50th percentile starter instead of last starter. Because I want a dominant team.

That's how I used to tweak the old Draft Dominator app (not possible in the current version I think); moving the baselines up to set the first player over $1 to be someone decent. The problem is that it felt like that method wound up overvaluing the middle of the curve, contributor-level players who

The following auction value methodology has been used for the past decade in my home league and wanted to share in case it provides value to others. Examples for my league are in parenthesis:

This is pretty similar to the setup I use, with one substantial difference. In my view, linear assignment of VORP points overvalues mediocre players and undervalues studs, partly because of the cost of roster spots, partly because of the higher reliability of projections for top players. My general rule of thumb in auctions is, if you're bidding on the best player left on the board, win the auction. Paying $24 for a $20 player is better than getting two $12 players at par value. I wanted my calculations to reflect that strategy, so I set up a sigmoid curve function.

The calculation is:

• VBDS = (VBD - (crossover)) / (sigmoid denominator) [In my league I start with 80, 20 but I tweak them]
• EXPZ = 1 / (1 + EXP (-VBDS)) [Sigmoid function based on VBDS. Equals zero when VBD = crossover].
• ValueS = (bid $ available) / (sum of EXPZ) * EXPZ

The crossover parameter sets the point at which the sigmoid curve crosses over from increasing to decreasing slope; I set it to a VBD number which represents, "after this, players get less interesting." In 2023 that was 50% of the top VBD number, which was guys like Rhamondre Stevenson and Jaylen Waddle. The calculation comes out to have those at about $10 of a $100 budget.

The sigmoid denominator sets the steepness of the curve. With it set at 20, the top projected player (which was Ekeler in our PPR league last year came out at $39. Set it to 15 and he's worth $41; set it to 25 and he's worth $37. I play around until I like the shape of the curve.

One thing to note when doing these calculations: The total bid $ available for a given player isn't "# of teams * budget". Assuming everyone costs at least $1, it's "# of teams * budget - number of player slots", because all teams have to keep enough money to fill their rosters.

Is the sigmoid denominator a percentage (0.2) or whole number (20)?

Whole number. In the spreadsheet it's =(E2-$C$446)/$C$447: E2 is the VBD projection for the player, C446 is the crossover, C447 is the sigmoid denominator. Basically changes the slope of the curve once you feed the result into the sigmoid function. Last year in my league the VBDS function ranged from +4.02 to -4.00. (For players with VBD<=0, VBDS = ((0 - crossover) / sigmoid denominator), so using 80 and 20 comes out to -4.00).

This is very helpful. I will try to implement within my own spreadsheet, thanks!

 

[Ray_T]

in the end, when bidding on a player your original prelim vallues can be somewhat meaningless. sometimes you need to go above your max value to get the player. but it is important to know when to walk away.

sometimes its easier to see where the value is once a couple of players have gone and you have to adjust your prelim values because they are a couple dollars too high (or too low)

usually I have found that if your prelim value for a top WR is off by 2 or 3 bucks, then its likely the same for most WR of comparable value. so you adjust your estimate and try to figure out which players become worth less. (if WR are undervalued, chances are another position is overvalued)

This is what a dynamic pricing model is good for. Set it up so that as soon as a player is won, values adjust for the rest based on points/position and such.

yeah but you need a certain amount of data first.

and not everyone has one of those.

what I make not of is that if the high end players go for more. ususally your mid range players will go for less. (as everyone has less cash when bidding on them) sometimes its by position. WR for more then RB for less(or something of the like) but usually its by value of the player.

either way it is important to see how the early auctions go. I usually throw out the first one or two as people are a little scared to spend money on the first player out of the gate. so sometimes you can get a bargain there.

 

[Scoresman]

I nominate players I want early and often.

Last year I was really targeting Kelce and my team strategy would change whether I got him or not. I nominated him first. Turns out he went for way more than I budgeted (happy for this in hindsight) but that info is good to know as early as possible because it affects your team strategy so much.

I’ve also learned not to really target one player like that.

 

[icehouse]

In keeper auctions, how much do y’all play keep away? Sometimes there are situations where some of the top players are being kept for low amounts, and if you release another top player that team can bid more than others and say walk away with 3 top players for like 50% of the budget.

 

[Instinctive]

in the end, when bidding on a player your original prelim vallues can be somewhat meaningless. sometimes you need to go above your max value to get the player. but it is important to know when to walk away.

sometimes its easier to see where the value is once a couple of players have gone and you have to adjust your prelim values because they are a couple dollars too high (or too low)

usually I have found that if your prelim value for a top WR is off by 2 or 3 bucks, then its likely the same for most WR of comparable value. so you adjust your estimate and try to figure out which players become worth less. (if WR are undervalued, chances are another position is overvalued)

This is what a dynamic pricing model is good for. Set it up so that as soon as a player is won, values adjust for the rest based on points/position and such.

yeah but you need a certain amount of data first.

and not everyone has one of those.

what I make not of is that if the high end players go for more. ususally your mid range players will go for less. (as everyone has less cash when bidding on them) sometimes its by position. WR for more then RB for less(or something of the like) but usually its by value of the player.

either way it is important to see how the early auctions go. I usually throw out the first one or two as people are a little scared to spend money on the first player out of the gate. so sometimes you can get a bargain there.

What I meant takes no historical data. Just a dynamic model that shows if more money was spent on the first player, or less money, vs what my value placed on that player was, it recalculates the remaining points of value vs remaining dollars to adjust to new prices.

 

[Payne]

I have done auction leagues for 25 years. Plenty of them.

Here is one thing that seems to be fairly consistent.

The first few nominees almost always go for more. I mean the first two or three.
Once that initial explosion stops, that is where the value lies.
If you have 5 WR's in your tier one, almost always the second or third of that group will be your best value.

SOMETIMES, the very first is the best value. If you trust your numbers and you have Lamb at 45 and he is at 38 0r 40, pounce on him.

I know this piece of advice has been explained over and over, but never be stuck bidding on a player who is the last of his tier. NEVER.

 

[Ray_T]

in the end, when bidding on a player your original prelim vallues can be somewhat meaningless. sometimes you need to go above your max value to get the player. but it is important to know when to walk away.

sometimes its easier to see where the value is once a couple of players have gone and you have to adjust your prelim values because they are a couple dollars too high (or too low)

usually I have found that if your prelim value for a top WR is off by 2 or 3 bucks, then its likely the same for most WR of comparable value. so you adjust your estimate and try to figure out which players become worth less. (if WR are undervalued, chances are another position is overvalued)

This is what a dynamic pricing model is good for. Set it up so that as soon as a player is won, values adjust for the rest based on points/position and such.

yeah but you need a certain amount of data first.

and not everyone has one of those.

what I make not of is that if the high end players go for more. ususally your mid range players will go for less. (as everyone has less cash when bidding on them) sometimes its by position. WR for more then RB for less(or something of the like) but usually its by value of the player.

either way it is important to see how the early auctions go. I usually throw out the first one or two as people are a little scared to spend money on the first player out of the gate. so sometimes you can get a bargain there.

What I meant takes no historical data. Just a dynamic model that shows if more money was spent on the first player, or less money, vs what my value placed on that player was, it recalculates the remaining points of value vs remaining dollars to adjust to new prices.

I get that. but most of us dont have a model like that at our fingertips.

I have no doubt that if you could run a calculation like that quickly it would be an advantage of sorts. I'm sure you could pay for one. wouldnt surprise me if FBG here even sells one.

but I have always been good at spotting patterns in the data so when I see this stuff(guys going for more than I predict) I file it away knowing there will be bargains later. its more or less that simple really.

 

[Scoresman]

I have done auction leagues for 25 years. Plenty of them.

Here is one thing that seems to be fairly consistent.

The first few nominees almost always go for more. I mean the first two or three.
Once that initial explosion stops, that is where the value lies.
If you have 5 WR's in your tier one, almost always the second or third of that group will be your best value.

SOMETIMES, the very first is the best value. If you trust your numbers and you have Lamb at 45 and he is at 38 0r 40, pounce on him.

I know this piece of advice has been explained over and over, but never be stuck bidding on a player who is the last of his tier. NEVER.

In my experience, this is highly dependent on who gets nominated first. I’ve noticed that if you nominate the clear cut top guys first, people will pounce and there will be next to no value. But if you nominate someone like a Derrick Henry or A.J Brown really early, people get shy, waiting on the super studs.

It again comes down to my number one piece of advice for auction leagues. Know your league and league mates’ tendencies.

 

[Pip's Invitation]

I'm too dumb to understand most of this stuff, but it does interest me and I think it's more or less what I'm doing with my prep (look at previous years projections then how much players went for). Still the number crunching is very interesting to me.

But like Mr. Pack said, auctions have a way of going sideways so my #1 auction strategy is have a plan of some sort. The most simple one I go with is "spend somewhere around half your budget for your top two guys who are ideally in the top 10-20ish ranked players" (for a 12 person league). Keep a little purse for those "middle round" guys, but don't be too stingy early. Tiers lists are much more valuable in auction to me than in snake, so have those ironed out. Give a rough estimate of how much you want to spend for each tier. And of course the golden rule of "know your league."

Another reason why keeping tiers is so important is that over the years I've found that the heaviest bidding wars are often over someone who is the last in their tier. So if you want someone from a certain tier, acquire them before the last one goes up for bid.

 

[Pip's Invitation]

I have done auction leagues for 25 years. Plenty of them.

Here is one thing that seems to be fairly consistent.

The first few nominees almost always go for more. I mean the first two or three.
Once that initial explosion stops, that is where the value lies.
If you have 5 WR's in your tier one, almost always the second or third of that group will be your best value.

SOMETIMES, the very first is the best value. If you trust your numbers and you have Lamb at 45 and he is at 38 0r 40, pounce on him.

I know this piece of advice has been explained over and over, but never be stuck bidding on a player who is the last of his tier. NEVER.

In my experience, this is highly dependent on who gets nominated first. I’ve noticed that if you nominate the clear cut top guys first, people will pounce and there will be next to no value. But if you nominate someone like a Derrick Henry or A.J Brown really early, people get shy, waiting on the super studs.

It again comes down to my number one piece of advice for auction leagues. Know your league and league mates’ tendencies.

Yep. In mine, often the first few nominations go for less than expected (which is how I got CMC for cheap enough to keep him this year), and then by the middle of the first round or so, people get more comfortable and the bids start going higher. Often the biggest overbids come in the middle of the auction when some people have too much money left and the talent on the board is dwindling.

 

[Donny Loved Bowling]

I really appreciate the OP for creating this thread, as well as everyone who has contributed to it. My primary league has been auction for over 10 years now, and while I have had some success, I've always felt that many of my league members hold an inherent advantage over me because of their mathematical prowess. (Several of them are engineers.) As I begin draft prep in earnest this week, I will attempt to use some of these methods to develop accurate rankings and auction values. In past years, I have sat at my draft table unable to overcome the gnawing feeling that many of the other guys in the room are far more prepared than me, despite my having pored over rosters and FBG articles for hours on end. The mathematical edge can be a powerful one, and very intimidating for those who know they don't have it. The step-by-step breakdown of your processes will hopefully allow me to better comprehend the overall statistical picture of my league auction, even if some of the jargon is a bit difficult for me. (You lost me at sigmoid curve.) Again, thanks to all who are contributing here, and please keep it coming.

 

[Instinctive]

in the end, when bidding on a player your original prelim vallues can be somewhat meaningless. sometimes you need to go above your max value to get the player. but it is important to know when to walk away.

sometimes its easier to see where the value is once a couple of players have gone and you have to adjust your prelim values because they are a couple dollars too high (or too low)

usually I have found that if your prelim value for a top WR is off by 2 or 3 bucks, then its likely the same for most WR of comparable value. so you adjust your estimate and try to figure out which players become worth less. (if WR are undervalued, chances are another position is overvalued)

This is what a dynamic pricing model is good for. Set it up so that as soon as a player is won, values adjust for the rest based on points/position and such.

yeah but you need a certain amount of data first.

and not everyone has one of those.

what I make not of is that if the high end players go for more. ususally your mid range players will go for less. (as everyone has less cash when bidding on them) sometimes its by position. WR for more then RB for less(or something of the like) but usually its by value of the player.

either way it is important to see how the early auctions go. I usually throw out the first one or two as people are a little scared to spend money on the first player out of the gate. so sometimes you can get a bargain there.

What I meant takes no historical data. Just a dynamic model that shows if more money was spent on the first player, or less money, vs what my value placed on that player was, it recalculates the remaining points of value vs remaining dollars to adjust to new prices.

I get that. but most of us dont have a model like that at our fingertips.

I have no doubt that if you could run a calculation like that quickly it would be an advantage of sorts. I'm sure you could pay for one. wouldnt surprise me if FBG here even sells one.

but I have always been good at spotting patterns in the data so when I see this stuff(guys going for more than I predict) I file it away knowing there will be bargains later. its more or less that simple really.

You can make it in excel in an hour or two. We're in a thread with people talking about creating pricing strategies and projections for values of all players using VBD baselines, so I kind of assumed people had excel and an hour or two to make it.

It's a huge advantage. Feel like it has always made me understand the value of tier-openers better and also the importance of tier closers and making sure you spend for value when it makes sense as things happen unexpectedly in a draft.

Edit: Honestly you could do it by hand pretty quickly live.

 

[Instinctive]

I have done auction leagues for 25 years. Plenty of them.

Here is one thing that seems to be fairly consistent.

The first few nominees almost always go for more. I mean the first two or three.
Once that initial explosion stops, that is where the value lies.
If you have 5 WR's in your tier one, almost always the second or third of that group will be your best value.

SOMETIMES, the very first is the best value. If you trust your numbers and you have Lamb at 45 and he is at 38 0r 40, pounce on him.

I know this piece of advice has been explained over and over, but never be stuck bidding on a player who is the last of his tier. NEVER.

In my experience, this is highly dependent on who gets nominated first. I’ve noticed that if you nominate the clear cut top guys first, people will pounce and there will be next to no value. But if you nominate someone like a Derrick Henry or A.J Brown really early, people get shy, waiting on the super studs.

It again comes down to my number one piece of advice for auction leagues. Know your league and league mates’ tendencies.

Yep. In mine, often the first few nominations go for less than expected (which is how I got CMC for cheap enough to keep him this year), and then by the middle of the first round or so, people get more comfortable and the bids start going higher. Often the biggest overbids come in the middle of the auction when some people have too much money left and the talent on the board is dwindling.

This is where the dynamic pricing helps - because some guys go for fewer $/point provided, there are more dollars left in the pool for fewer points. So supply of points is going down while supply of dollars is going up, relatively speaking. It creates an inflation effect for the cost of a point.

 

[Pip's Invitation]

I have done auction leagues for 25 years. Plenty of them.

Here is one thing that seems to be fairly consistent.

The first few nominees almost always go for more. I mean the first two or three.
Once that initial explosion stops, that is where the value lies.
If you have 5 WR's in your tier one, almost always the second or third of that group will be your best value.

SOMETIMES, the very first is the best value. If you trust your numbers and you have Lamb at 45 and he is at 38 0r 40, pounce on him.

I know this piece of advice has been explained over and over, but never be stuck bidding on a player who is the last of his tier. NEVER.

In my experience, this is highly dependent on who gets nominated first. I’ve noticed that if you nominate the clear cut top guys first, people will pounce and there will be next to no value. But if you nominate someone like a Derrick Henry or A.J Brown really early, people get shy, waiting on the super studs.

It again comes down to my number one piece of advice for auction leagues. Know your league and league mates’ tendencies.

Yep. In mine, often the first few nominations go for less than expected (which is how I got CMC for cheap enough to keep him this year), and then by the middle of the first round or so, people get more comfortable and the bids start going higher. Often the biggest overbids come in the middle of the auction when some people have too much money left and the talent on the board is dwindling.

This is where the dynamic pricing helps - because some guys go for fewer $/point provided, there are more dollars left in the pool for fewer points. So supply of points is going down while supply of dollars is going up, relatively speaking. It creates an inflation effect for the cost of a point.

Yep. Not being a math or spreadsheets guy, I deal with this by acquiring as many people as I can at prices I think are reasonable early on. This allows me to sit out the inflationary part of the auction but, if I've done things right, save enough money to get some of my sleepers at the end.

 

[BroncoFreak_2K3]

The Fantasy Pros Draft Assistant with sync app I use has a button you can click that adds in an inflation element as the auction progresses. I’m interested to see how this works as this is the first year using it with an auction draft.

 

[Statorama]

 

[dkp993]

How would you compensate for keepers with this process? They are kinda outside of the auction process somewhat as we load those in pre-draft. However, their keeper values are subtracted from each team's total auction budgets.

For my auction league, keepers are kept for their previous year's auction value * 25% inflation annually and can be kept by the same team up to 2 consecutive years.

I have never looked at keepers, but I will take a look this week to see if the formulas can be adjusted in some manner to compensate.

I’ve been running a sheet for my keeper auction for the last 12-+ years and have all the historical data as well.

I use a very similar format as azmat with the exception that I use yearly points (not ppg) to calculate my VORP from the first 1$ player at the position (I run 2,5 and 10 year averages to help spot trends and making a necessary adjustments to who I think will be the first one dollar player this year at that position).

To adjust for keepers I keep a separate spreadsheet that is linked to my auction sheet that tracks the keepers and subtracts the dollar amount kept for those players from the remaining auction dollars left to use. I also subtract the players from the projection sheets that are kept and their corresponding projections thus the VORP stays true. Doing this gets me a cost per point, which I then turn into an auction value and compare data for the positions. It’s been scary accurate over the last five or six years and I’ve been pretty good at being able to predict what the 13th running back off the board will go for (or 5th or 11th, etc etc). And while there are certainly outliers as auctions are unpredictable and people fall in love with players for different reasons, the averages across the spectrum give me a massive advantage over my league mates imo (I’ve won 6 league titles in the past 10yrs since I started incorporating this method).