Hoop Pointers: Price Fixing, The Sequel
February 6, 1998
I've been delinquent in following up on the Price Fixing article I wrote way back in late December. I did receive seven responses with ideas, and several of them suggested very similar changes which were also in line with my thoughts. I did discuss this briefly with Erik at Smallworld last month, and he seemed receptive to considering the ideas for the upcoming baseball season.
As a brief recap, the basic problem is that if a player is priced too low at the beginning of the season, he is on so many initial rosters that there aren't that many managers left to buy him. Since only trading activity influences prices (and not the initial drafting), this dampens the potential for these player's prices to increase to a level which is more consistent with similar players. Also (and perhaps more importantly), players who seem overpriced at the outset aren't drafted by many teams, and thus there is little ability for their prices to fall to a fair level, since no one has them to sell.
The best solution seems to be along the following lines. In addition to current practice of adjusting prices based on the relative buys and sells, there should also be an adjustment which compensates for the degree of current ownership. For example, if few teams own Mookie Blaylock, and no one is buying him, this implicitly indicates that his price is too high. In effect, for a player who is lightly owned, a "don't buy" decision is the virtual equivalent of a "sell". This player's price should gradually drop until it reaches a point at which managers begin buying. Similarly, for players who are widely owned, a "don't sell" decision is the virtual equivalent of a "buy". Prices for these players should gradually rise until they reach a level that prompts selling.
The significant benefit of this approach would be to have a process by which each player's price will necessarily trend toward a level which produces average ownership. No longer should there be players who are persistently cheap - and who therefore show up on an overwhelming majority of rosters. And no longer should there be players who are persistently "out of play", since their prices will keep falling until they reach a level that induces some managers to buy them.
Calibrating this feature would require some inside information which only Smallworld has access to. But it would seem that each player's ownership percentage (that is, the percentage of teams which own the player) could be compared to the average, and a "drift factor" could be added to the normal "buy-sell" factor to determine the periodic price change. For players whose ownership percentage is near the average, the drift factor would be zero, and the process would work as it does today. As the ownership percentage increases above the average, the drift factor might also increase, so that someone who is owned at 150% of the average might have a smaller upward drift than someone who is owned at 200% of the average. The reverse would happen for players whose ownership is below average, although the scaling might be different.
There is a related issue that may need to be addressed in order to prevent unintended consequences of this type of change. At the beginning of the season, it appears that prices are set at a level such that the average player price is roughly equal to the average roster capacity. For example, if you have to buy 12 players with $50 million, the average price you can pay for each player is $3.3 million. If the average price of all players is $3.3 million, then you would expect an efficiently priced field would result in each player being owned to the same degree. However, as the season progresses, average roster values tend to inflate. Although the average player price today has only drifted up to approximately $3.5 million, I would guess that the average roster value is $60 million or more - at least for teams which have remained somewhat active. If so, then the average roster slot today would support a player worth $5 million. In this instance, you would expect higher priced players to have higher ownership percentages than lower priced players. If the drift factor doesn't compensate for this, you could conceivably find that higher priced players would tend to perpetually drift upward (since they would always tend to be owned more than average), while lower priced players would have a persistent downward bias. Perhaps the drift adjustment on any given date should only reflect a player's ownership percentage relative to the average for other players of comparable value. So, you would compare the ownership percentage of a $5 million player against the average for other players worth roughly $5 million. Or, more broadly, against the universe of all players worth $1.6 million and above (which today happens to be the subset which produces an average price of $5 million per player.) Maybe this is splitting hairs, but I expect these hairs are worth splitting.
Well, enough of this. The ball is now in Smallworld's court. I'm sure they'll make some intelligent changes.
Return to Guru's Hoop Pointers index
Hoop Pointers is written by Dave Hall (a.k.a. the Guru), an avid fantasy sports player. He is not an employee of any of the fantasy games discussed within this site, and any opinions expressed are solely his own. Questions or comments are welcome, and should be emailed to Guru<email@example.com>.