Subject: RotoHog Salary Cap baseball|
Posted by: Guru
-  Sat, Mar 13, 2010, 12:04
Probably worth setting up a separate thread for this.
The RH Salary Cap game is essentially a clone of the TSN format. The overall prizes are lower, but the division winner gets $60.
I haven't yet studied all of the rules, but it appears that the 5-day pitcher repricing may not be used (I didn't see it mentioned anywhere - CMIIW), so Randro strategies may be more useful this year, and relief pitcher trains may be less impactful.
I'll plan to offer the same tools for this game (sortable stats, Assimilator, etc.) that I offered for the TSN games. I'll try to have preliminary versions up within the next week.
Team prices are mistated on the front page. A 3-pack costs only $35 (earlybird), and a 5-pack is $50. I assume they'll figure that out eventually. The ordering page has the correct prices, which were the same prices typically used for TSN. (Of course, you can't use TSN prize credits to buy a team.)
If you notice other differences from prior TSN games, please mention them here.
Sat, Mar 13, 2010, 12:09
If the higher pricing on the front page is what they really intended and they are simply charging thw incorrect prices on the order page, then it probably behooves you to get your team(s) now. I notice that the official game rules also mention the higher prices ($55 for 3, $80 for 5 teams).
Sat, Mar 13, 2010, 12:12
The prize rules state this with regard to premier league prizes:|
League prizes for the top 2 teams with the highest overall point totals amongst all paid teams in leagues with eight (8) or more paid teams (Per League)However, the prize table lists only one league prize ($60). I assume the mention of a second league prize is an error, although I have no idea.
Sat, Mar 13, 2010, 12:13
Finally, if there is enough interest, I'll set up a separate league just for GuruPatrons. If you are interested in joining, please post your interest here. I'll defer on setting that up until I see that we have enough interest.
Sat, Mar 13, 2010, 14:01
I'm interested in gurupatrons
Sat, Mar 13, 2010, 14:33
Yes, although the rules say that the premier game has the same formula as TSN but the basic game is at 50%, both games list points for last year at the 50% level. Doesn't really matter, but they ought to get that right.|
The 2009 points listed for closers are all missing the points for saves.
Quality control appears to be a real concern.
Sun, Mar 14, 2010, 15:30
With the 5 day pitcher cycle and WINS and Saves with the same value ( 15 points ) for premier, Rotohog is a copy of TSN Salary Cap.|
I already set up a team in basic to see how it behave and well it is a complete new look and feel for buying players, looking stats, etc.
Fri, Mar 19, 2010, 00:37
Interested in Guru Patrons.
Mon, Mar 29, 2010, 10:34
Still looking for interest in a GuruPatrons division. So far, I have not set it up as there is insufficient interest expressed here.
Mon, Mar 29, 2010, 11:25
I am interested.
Thu, Apr 01, 2010, 12:25
Copied from today's blurb:|
It was pointed out to me this morning that the scoring formula for the RotoHog Salary Cap game was revised sometime after the initial launch. Originally, all scoring points were exactly halved from the TSN formula of recent years. But after that, the points for saves were cut in half again. RotoHog’s rationale for that seems to be flawed relative to the structure of the game, but then, no one asked me. The consequence is that most closers become irrelevant, since all pitching spots are available for any pitcher, and closer prices were never readjusted to compensate for the formula change. The highest ranking closer in total points for 2009 was Broxton, but that now ranks behind about 35 starting pitchers. So it really makes no sense to own a closer unless you find one that is grossly underpriced. And since all pitchers are grossly underpriced to begin with, it is inefficient to go dirt cheap on pitching.
RotoHog may be able to manage the software elements efficiently, but they lost a lot of institutional memory and understanding related to game design when they took over the game without any support from TSN. Too bad.
My sortable stats still reflect the higher points for saves. I probably won't adjust that for 2009, but 2010 will use the correct formula, of course. Suffice it so say that you probably shouldn't be looking at closers anyway. Starting pitchers offer much better value.
Fri, Apr 02, 2010, 10:52
Still only 5 have indicated interest in a GuruPatrons division. Not enough to follow through.
|16||Dave R |
Fri, Apr 02, 2010, 13:06
I'd be in
Fri, Apr 02, 2010, 15:10
You were one of the 5
Sat, Apr 03, 2010, 22:27
I'd be in if needed but it seems we are more than 1 away.
Sun, Apr 04, 2010, 13:27
Went to set a lineup just now, and was greeted by "500 Internal Server Error".|
Some things never change, regardless of who's running the game.
Sun, Apr 04, 2010, 19:21
Is anyone else able to access their Premier Salary Cap team? I've been getting a "500 Internal Server Error" all day, and I just got an email response from Customer Service that suggests they are unaware of a general problem.
Sun, Apr 04, 2010, 21:07
I'm have no trouble with access.
Sun, Apr 04, 2010, 22:39
Got an email from RH customer service saying that they found the problem and corrected it.
Tue, Apr 13, 2010, 11:48
I teased you in yesterday's blurb yesterday that I’d comment on the RotoHog Salary Cap games today. So I will.|
While the game format is generally the same as the previous TSN Salary Cap games – in scoring, trade capacity, roster size, etc. – the one major change is the mechanics of the daily price change formula. I’ll focus on the premier game in this discussion, which corresponds to the TSN Ultimate game – but the comments are generally applicable to the basic game as well.
First, some background. The TSN daily price change formula depended on a day’s trading activity. If a player enjoyed net buys, his price would rise, and vice versa. The amount of the change depended on the percentage of buys and sells that a player received vs the overall total for all players for that day. The formula was not linear, and the biggest gainers and losers on any given day gereally topped out at a price change of around $200K. And there were very few players each day with a change anywhere close to that.
A consequence of that formula is that team roster values tended to inflate, but only at a rate of several hundred $thousands per day. A big day of gainers produced a roster gain of a $million or so, but those magnitudes would typically not persist for long. All teams started the season with a value of $50mil. By the end of April, the wealthiest teams might have appreciated into the low $60 million range. By the All Star break, the richest teams were in the mid-$80s, and some teams might be touching $90m. With a 14 player roster, then the richest teams could support an average player price of about $4.5m in early May, and $6m in July. The most expensive players were priced around or above $10m.
After about a week this season, the vast majority of teams already have a team value in the mid $70s, and most teams are gaining close to $3m per day. At this rate, teams will crossing though the $100m threshold by the end of next week. And by then, the most expensive players may have dropped below $10m. (Pujols is currently priced at $10.9m, and is dropping $.24 mil each day.)
It seems likely that team values will grow so large, so fast, that pretty soon, this will no longer be a salary cap format, since player price will not longer be a limiting factor. But rather than speculating on the future of the game, I want to instead discuss the mechanics of the RotoHog price change approach, since understanding that is necessary to understand how to manage a team today, next week, and next month.
Stay tuned... Time to take a breather.
Tue, Apr 13, 2010, 12:20
Before continuing, props to several people at the TSN baseball forum who did most of the research to figure this out.|
Let me use this segment to describe the price change method being used. I'll defer comments on the implications until later.
The formula assumes that there should be a direct relationship between a player's ownership level and his price. Highly owned players should have high prices. Low prices should be associated with lightly owned players. To the extent that a player's actual price differs from the theoretical (target) price, his price will move toward the target. The maximum daily change is $250K (in either direction). As the current price approaches the target price, the amount of the daily change is reduced. I haven't yet figured out exactly how that daily movement is determined, but for now, that's a detail that isn't particularly critical. Eventually, we'll probably figure that out, too - if it ever does become important.
Perhaps this can best be understood visually. We can begin to map the relationship between ownership and target price if we look for players whose daily price change is very small. The following graph plots price vs ownership for players with a daily change yesterday of $20K or less.
Currently, there are enough data points at lower prices to get a pretty good sense of the curve. As prices and ownership rise above $5m and 5%, we still lack much direct empirical data. The point in the upper right is Ryan Howard, whose price fell $20K yesterday, so his target price should actually be slightly to the left. The next highest data point is Jason Bartlett, whose price rose $20K yesterday, so his target price should actually be slight to the right.
There are a number of players at much higher ownership levels. In fact, 24 hitters are on more than 10% of all teams, and the most heavily owned player, Jason Heyward, is on 43% of all rosters. Almost all of those players are still appreciating more than $200K per day (five at the max of $250K), so it's tough to get a sense as to where those prices might find equilibrium. Joe Mauer is on 11.6% of teams, and his current price ($7.3m) gained only $130K yesterday, so he's honing in - but it's still tough to say.
What do we learn from this table? If you know a player's ownership level (and that data is readily available at the game site), then you can pretty easily predict what will happen to his price. If you are looking for the biggest prospective gainers, simply find those players whose ownership level matches a target price much higher than their current price. So far, those players are plentiful.
For pitchers, the graph isn't quite as tight, but that is apparently due to some averaging of the ownership levels. RH has said that they are taking into account 5-day pitching cycles, so I assume from this comment that the ownership levels for determining target prices are based on the average over the previous five days. I haven't rigidly tested that assumption, but it does pass a general "sniff test" on a few data points.
For the basic game, the graph is very similar in both magnitude and slope.
So what is likely to be happening for ownership levels above 10%?
Visually, the relationship appears to be curved and not linear, and with so few points at the higher levels to work with, we are generally left to guess. Suffice it to say that for someone like Jason Heyward, who is currently priced at $3.04m, he has a long way to go - even if his ownership stays perfectly flat. In fact, even if his current ownership were to be cut in half!
Tue, Apr 13, 2010, 12:59
Is this a sensible model?|
The model presumes that ownership level and price should be related. If Pujols is the most valuable (i.e., the most productive) hitter, then most would be comfortable with the premise that his price should be the highest of all hitters. But does it follow that he should also be the most highly owned hitter?
Obviously, at the start of the season, the answer is "no." When teams start with limited funds - and Pujols would consume almost 25% of a team's initial salary cap space - then it's unlikely that he will be very highly owned. And, of course, he wasn't. So while his opening price of $12.8m might have been justifiable on a pure production-based model, the model being used determined that his target price - based on lower ownership - should be much lower.
Of course, once teams gain enough cap space, they'll be able to afford him. And if he does turn out to be the best hitter, his ownership will climb, pushing his price higher. So arguably, although the start of the season creates an apparent anomaly, one might argue that the model will work as the season progresses.
Even so, there is still flaw. There are only two roster spots that Pujols can populate - either 1B or DH. If inflating team wealth means that salary cap becomes no constraint, then there will still be little demand for the 3rd or 4th or 5th best first baseman. Teams with adequate wealth will simply own the top player at each position. True, there may be some shuffling due to favorable matchups or hot streaks, but with trades limited in number to 3 per week, that's unlikely to be a material strategy.
So perhaps the issue becomes one of whether enough teams can amass enough wealth to avoid any effective constraints. At this point, that seems plausible. I sampled a number of teams who were ranked around the 75%-ile in points, and they all had amassed team values in the mid-$70s. Teams ranks near the median are almost as wealthy - and at the current pace, are only a day or two behind the wealthiest teams. I expect that this rising tide is likely to float all boats - or at least enough boats to matter.
So it's not yet clear to me that this model makes theoretical sense, even after we pass the first few weeks. (I'll admit, though, that there are some complex interactions that make it hard to project ahead with complete confidence.)
Nonetheless, in the first week, this approach clearly makes no sense. And the possible implication is that this game will cease to be an effective salary cap format very soon - if it ever was to begin with.
Tue, Apr 13, 2010, 13:45
Guru you are right on- The highest possible salary cost now stands at just over 109M which at a rate of a mere 3M per day even my team will reach in 12-13 days. Effectively by the end of April those who have saved hitter trades will all own the 9 most expensive hitters.|
Mauer 7.30, Pujols 10.93, Utley 8.29, ARod 8.73, Henley 9.47, Braun 8.38, Holliday 7.54, Cruz 6.60, DH- Fielder 7.92 Hitter total 75.16 top 5 pitchers are 34.23
Looks like we all got ripped off if we paid for a salary cap game.
Tue, Apr 13, 2010, 13:46
Nice work Guru.|
Teams with 70 million after two weeks is rediculous.
Im glad I didnt play.
In addition, when the formula becomes more widely known it would seem a fairly simple task to have all of the major price gainers each day.
Tue, Apr 13, 2010, 14:05
Actually it is hard not to be gaining at least 2.5M per day as everyone you can afford is gaining at least 200K a day and the only ones that are dropping are the high price performers like Pujols. I guess the only decision left is do you wait the 12 days so that you can afford to buy all 9 of the top players at once or do you slot them in as you can afford them in which case |
since you are sacrificng about 500K a day each it could take you an extra 2 weeks to afford all of them.
I don't know what the solution is- maybe make this a 6 week game ending Memorial day and starting over with a new free team with a new pricing logarhythm June 1? Since this is a Rothog problem methinks Rotohog should pick up the tab for the multiple prizes that would require ? LOL.
|29|| Kal Zakath|
Tue, Apr 13, 2010, 14:46
Here's what I am seeing in the future, though....|
We're about to hit the end of these incredible price increases on the team scale. One by one, the players are reaching their Equilibrium Point (EP), and are no longer gaining. When this is happening with productive players, you need more buys to drive their prices up higher.
Within the next few days, I'm expecting that virtually everyone under about $6-7 million will be at that EP, and only the most owned players will be going up after that.
The next thing that comes up will be the sell offs and buys for the guys coming up. What we're seeing right now is 50-90k spikes and drops based on a single owner move (the variability seems that sometimes it jumps over or under the EP with the previous price update).
I'm forseeing a situation with hitter trades where there are moves made on schedule much like with the Ultimate hoops game - where there are going to be significant ownership moves on the studs based on an extra game or two - Pujols (once he finally bottoms out and gets ownership to support price) going to SD, SF, and LA while Howard is at home with 1 or 2 extra games? The move will be made. Pujols back at home? Howard -> Pujols again. With the large price swings possible with just a few managers buying or selling a player, it could be a couple million being moved around quickly.
I'm not sure we're going to see the huge huge numbers that are being discussed here unless people actually try for them. The question will then be, do you actually need to try for them. Is just picking up the studs as they get to the EP point going to be enough, and then holding them just avoiding the losses going to be enough?
I see with this system, even though everyone has a ton of cash, less, not more variability in rosters, which is something that plagued the TSN game as seasons were played. When everyone has Pujols, it takes him out of the game for the most part. When everyone has Utley, Ramirez, A-Rod, Braun, etc., the same.
Tue, Apr 13, 2010, 16:11
I was no great fan of the TSN system either. It had its flaws as well. So frankly, I'm glad to see something different tried. This might not be the answer - but it might be possible to use it as a basis with other modifications.|
At some point before too long, Pujols will not only reach his price trough, but at that point buying will begin, which will start to push his price back up. And since people can afford him, they'll flock to get him, and the inflation will push him back to $12m or higher. (We don't know whether there is a max price or not.)
In the TSN game, some players were always chronically cheap, and those players tended to appear on most competitive rosters for the better part of the season. In this format, chronic cheapness won't happen, because anyone who shows up on a lot of teams will gradually (or not so gradually) become expensive.
I suspect pitcher rotation may become a very important strategy this year. Not for the price action, but to maximize points through judicious selection of good matchups. Once Pujols (or the like) gets locked into a lineup, I think it would take quite a schedule disparity to prompt much turnover. Maybe an impending road trip to a series of tough hitters parks would do it. But if you put together a list of the top player at each position, most of those guys can hit anyone, anywhere.
|31||Kal Zakath |
Tue, Apr 13, 2010, 17:15
True, but then it becomes discipline to not use all those hitter trades you're going to be getting.|
Could you really hold Pujols through what was going to look like 2-3 million in losses over a week or more if he weren't hitting a ton and was on a long road trip and someone else was red hot?
Same with Utley, Ramirez, Rodriguez, Mauer, etc.
Those trades are going to be used by people - the only question in my mind right now is when. A hot Pujols doesn't get traded if he's playing in the Grand Canyon for a week. A colder one does if one of the other big 1Bmen is hot. I'm really not sure if just hot or not is enough to make people move, but I think that people will outthink themselves in scheduling - that will be the last spot where you're going to be able to make up ground if everyone has the same players as it seems we are destined to have.
Tue, Apr 13, 2010, 18:44
You say the game is going to be the game is going to be a game of skill where people leave pujols because he is going away or jump on howard because he is coming home and ride hot streaks. But is this skill or luck? Is there a formula to know when how streaks are coming? I know some inferences can be made from the most extreme pitcher parks like coors and petco but are the best hitters that impacted by the things in between? To use your example of selling pujols when he goes away, last year Pujols had more runs, more rbis, and more HR with a better BA on the road with even PA. he was below average in both petco and coors (with very small sample sizes...). Of division teams where sample was a bit bigger, he was bad at the reds who have bad pitching and good ballpark so he should do well. At hou same thing.|
2008 he was better at home, 2007 on the road (significantly), 2006 better at home.
Are their trends or is there luck? I was thinking id look into more players but did more details on pujols. Does 3 straight 0-4s mean he will suck next time? Or is he just as liking to hit 2 HRs that game?
Thu, Apr 15, 2010, 21:14
The general concept of this model is fine. If it weren't for the magnitude of the changes this wouldn't be as bad as when they changed to 5-day pitcher repricing. I am probably in the minority but I didn't like that and I think that was one of the worst changes they ever made and one of the many things that led to the downfall of the SN game. But people loved all that RV which is why I figure most people are loving Rotohog right now. Obviously the magnitude of the price changes needs to be tweaked. It's a long season. A star player is going to get injured and be out a month, and in that month his price will drop $7 million. If this model had been in place last year Mauer would have missed the first month and then returned at $500,000.
Fri, Apr 23, 2010, 09:05
Hoo boy... Posted at the RotoHog forum:|
The RotoHog.com team continues to strive to provide the greatest fantasy sports games ever built. We feel it is necessary to not only listen to the requests of our users, but actively implement features which will speak to our pledge to provide the best game experience for our users.
On April 25th, we will be modifying our Salary Cap pricing system to be more responsive to the daily activities of the trading floor. Starting Monday night, player prices will fluctuate more based on the daily buys and sells than of our previous pricing system.
We will continue to make improvements and enhancements to our fantasy products as we strive to pride a superior fantasy product.
Please continue to contact us with your feedback, questions and concerns.
The RotoHog.com Team
Sat, Apr 24, 2010, 01:06
I was bummed when I decided to pass on baseball this year. That lasted for a few days once I saw the pricing structure. Guru - I agree it's nice to have something different, but it is a huge joke the way they have it.
|36||Dave R |
Mon, Apr 26, 2010, 07:33
I guess I should read more carefully, I interpreted the pricing mechanism to be changing yesterday, but it looks like tonight's the night.|
Anyone playing got a feel for what might happen? I have a guy like Prado, who's been +$250K nearly every day this year, inflating from a season opening price of $1.66M to last night ending at $6.83M. ( I wish my 401K portfolio behaved the same ;)
With ownership decreasing from 183 to 154 over the past few days, I would think he will be a prime example of what will occur under the new system tonight.
Anyone holding to see what happens, or are people bailing on the Prado's of the world?
Mon, Apr 26, 2010, 09:47
I think you read it correctly. In fact, they clarified their initial statement by saying "The Salary Cap pricing system update has been postponed until the evening of April 25th."|
I don't think it happened, though.
Regardless, I'm waiting until I see something to react to. Prado would seem to be vulnerable. Also Heyward. But we don't know whether this will be a system, based mostly on trade flows, or only partially - and if the latter, how that balance will be struck.
Nevertheless, it's hard to argue that guys like Prado and Heyward offer good stats-based value right now.
|38||Dave R |
Mon, Apr 26, 2010, 10:01
This is sure confusing. If you read the post above, it says starting Monday night , which I'm sure is what I saw this morning, leading to my post.|
However, on their site now:
On April 25th, we will be modifying our Salary Cap pricing system to be more responsive to the daily activities of the trading floor. Starting Friday night, player prices will fluctuate more based on the daily buys and sells than of our previous pricing system.
So are they giving the new pricing a weeks worth of buy and sells?
Mon, Apr 26, 2010, 10:33
So maybe the reference to Sunday night is when the new tracking of buys/sells begins?|
Their original message said that they were starting this on Friday, April 23. They subsequently changed it to read Sunday - but never changed the reference to Friday. So I think that is just a proofreading error.
Quality control in communications is not their strong suit. There are still errors in their rules pages (e.g., incorrect scoring formulas, incorrect prize specifications) that I have pointed out to them, and have been acknowledged by them, but never corrected.
This seems to be an operation that lives by the credo "Listen to what we mean, not what we say."
|40||Dave R |
Tue, Apr 27, 2010, 06:55
Ouch, seems like a " wait and see " approach was costly. There were to many losers to avoid anyway, at least most contributed positive points. |
Toss in a couple pitchers that lost $, including Lincecum who pitches tonight
Tue, Apr 27, 2010, 12:55
Price sensitivity looks pretty leveraged. Did just a few quick lookups:|
Cantu, -$230, ownership dropped by 14
Prado, -$180, ownership dropped by 6
Mauer, -$90, ownership dropped by 3
Choo, -$60, ownership dropped by 1
If these are going to be symptomatic, then one team can drive a pretty significant price move for any given player.
Wed, Apr 28, 2010, 10:10
Once again, very highly leveraged.|
Hitters with only one net buy generally went up $50K. Three net buys gained $100K. There is clearly another component - probably some remnant of the prior ownership model - but as advertised, daily trades are the dominant factor.
Hitters with only one net sell dropped by a similar amount: -$60K. Just 4 sells was enough to send a price down by $140K.
On the pitcher side, it is obvious that some sort of multi-day trade period is operative. My best guess is that a 5-day period is being used, but starting with April 25 (Sunday), as that seems to produce a much better correlation with observed price changes than going back a full 5 days. If that's the case, I'd expect that 5 days will be operative once we get to Friday and beyond. However, there are still some anomalies aren't explained by using trades over either period, so it's probably going to take some more time to sort this out for pitchers.
A few pitcher examples:
price loss is -$120K
Ownership change in last 5 days: +8
Ownership change since Sunday: -60
Total ownership: 442
He's highly owned, and with net buys in the last 5 days, his price seems like it should be moving up. This leads me to suspect that the full five days aren't being used - at least not yet.
price gain is +$90K
Ownership change in last 5 days: -43
Ownership change since Sunday: +20
Total ownership: 308
Once again, this seems to support a less-than five day period.
price loss is -$80K
Ownership change in last 5 days: -2
Ownership change since Sunday: +5
Total ownership: 44
This one is hard to explain based on my Sunday hypothesis. So something else is at work.
Wed, Apr 28, 2010, 13:06
The Price price change would make sense if the two components are 5 day net buys as well as overall ownership.|
The low ownership isnt enough to offset 5 net buys.
Wed, Apr 28, 2010, 13:32
Actually, ownership of 44 is pretty high in this game (5%).|
Regardless, if that's the explanation, then how do you rationalize Wainwright's loss? He has very high ownership (51%), and +8 buys in the last 5 days.
Wed, Apr 28, 2010, 14:18
If anyone wants to do some more sleuthing and/or attempt some reverse engineering, I have been capturing raw data on ownership and price changes since mid-April for the premier game. |
Here is the link for April 25 data (run at 8:15am ET on the morning of April 25):
For other dates, just adjust the date portion of the URL accordingly.
I don't have ownership change directly, but you can calculate that by comparing the ownership for two dates.
|Rate this thread:||If you wish, you may rate this thread on scale of 1-5. Ratings should indicate how valuable or interesting you believe this thread would be to other users of this forum. A '5' means that this thread is a 'must read'. A '1' means that this is a complete waste of time.|
If you have previously rated this thread, rating it again will delete your previous rating.
If you do not want to rate this thread, but want to see how others have rated it, then click the button without entering a rating, or else click here.
Post a reply to this message: