Despite the fact that we were the first to bring out a computerized valuation
system, we are strongly against computers in the draft. Also note that if you
have a partner we don't really object to a second person tracking information
on the computer but we still believe that the primary drafter should not have a
computer in front of him.
Our concerns with a computer in the draft basically focus on 2 things: slower access to information and lack of "artificial intelligence".
First, we are hearing that a lot of leagues are "outlawing" computers in the draft so it makes little sense to devise a system that will soon be illegal. The reason most computers are banned are that they (choose all that apply to your league): a) slow down the proceedings; b) give a perceived advantage to the user; c) give a perceived advantage to the owner who can afford a laptop; or d) they take up too much room and infringe on everyone else's space.
Also, your laptop will have a very small screen and you cannot get a lot of information on one screen causing constant scrolling. Actually in a recent review of one of our competitors products, the reviewer describes how his program crashed 1/2 way through the draft leaving him completely in the dark, as it were. A major problem there. Whether something happens with our program, your computer or Windows locks up or crashes (unheard of - Bill Gates is shocked) your draft and therefore your season is ruined.
Next, regarding artificial intelligence. The computer you bring into the draft looks to a pre-programmed set of facts and assumptions. It cannot replace human judgment in the draft. We liken the draft to a game of poker, sure it is nice to know the "odds" of the cards in your hands, but it is equally, if not more important, to be able to observe the other owners. Who is biting their lip or shaking their leg. You miss this buried behind a computer screen. The computer lacks true artificial intelligence and this means you might as well pre-program a robot to do your draft. No thinking allowed!
One ironic twist involves a program for fantasy baseball leagues. One of our
former baseball customers designed a program for John Benson to use as
"his" draft software. However, shortly before the introduction of his
system, John Benson wrote an article in
The other, and bigger problem, is speed and access to information. As noted above, most people with computers will have laptops. If it is not a Pentium III you are in a hole to begin with. Then, you have the size of the screen. In our league there will be about 110 spots open. All the players who should be taken will fit on two pages. One baseball program that we saw only shows 14 players on a PC. Thus, it would take 8 view screens to see all the players! Also, our system is very complicated and comprehensive. Even a Pentium III chip may take 7 to 10 seconds to recompute every player's value. Also, as players are taken some go over and some go under so by the end of the draft there is very little deviation from the initial numbers.
Contrary to what you may think it takes longer to find a player on the computer. If you have our alphabetical list of all players handy, pick it up. Now find Eugen Kingsale or Randy Johnson. One system we have seen would have you spell his name, and close doesn't count. Good luck. Another baseball system has a list of all players alphabetically like ours does. Still, you have type in the first letter, then scroll down the screen to highlight, then hit enter and then wait for the computer to pull him up. The third way used by baseball systems is to find him by position by value which is great if you know his position and value. But if you do, why do you need the system?
Believe us, by the time you find a particular guy, the bidding may have moved past you. You can find the player on the overall alphabetical list faster than on the computer and be crossing him off each of the lists as noted in the instructions while the bidding is moving around the table.
Even on a computer screen, you cannot use the scroll function faster than your eye can see it. In other words, it saves no time to scroll with a cursor or your finger. Actually, since you can reverse direction faster, it is quicker with your finger. When you do find the player on our list, it has the positions he plays, projected stats and value. Once you find him on the computer, you have to tell the computer to "call him up". Thus another delay.
The other main place the computer slows you down is when you are in the bidding with 2 or 3 other teams. If you have a laptop you will be constantly calling up one team or another to recheck their rosters. You will not be looking at their reaction to the bidding. If you have the rosters on paper you can see multiple teams at a glance and spend your time focusing on what is now important.
Also, once the player is taken it would seem that the computer could update the available salary and maximum bid faster than you can by hand. However, usually it cannot. Here is why. Player A is chosen by team 6. On the computer you first hunt down the key to assign a player to a team. Next you have to find and highlight that team. Now you have to assign that player to a position on that team through key strokes. Finally you have to type in the salary, but do it carefully since you may never catch an error until it is too late. By the time you do that you could have written the player on the team with his salary at his correct position. You would also have time to update that team's available salary and maximum bid.
Thus, with a computer, it takes longer to find a player, see rosters in your league and due to the number of key strokes involved it also takes longer to assign the chosen player to a team and update that team's available salary!
We have seen computers in use in the LABR drafts for the past 7 years. Virtually everyone using a computer takes longer to find players and after the player is chosen it is the computer users who are the last ones still entering information.
Another big concern we have is "experts" who are marketing a system for use in the draft with pre-set targets for categories. Essentially what this assumes is that you are drafting for balance. But if your league allows trading, draft for value and trade for balance.
In most drafts you have 2 options. First, filling your needs no matter the
cost, or second "over-acquiring" a category cheaply. We prefer the
latter strategy. Assume there is a run on power hitters in your draft. You
could go out and pay 35 or 40 for Barry Bonds or Manny Ramirez. On the other
hand you may get a decent base stealer like Ramon Santiago or Tony Womack for 5
or 14 (which is what they respectively went for in the 2003 LABR League draft).
During the season teams will start to fade and they will want to trade their
power for keepers for the next year. Thus
One last item that we have seen is to use the computer to calculate the change in the draft inflation during the draft. Two main problems with this. First, there is generally little change during the draft until the end. Remember if Barry Bonds has a true value of 45 and a draft value of 54 if he goes for 55 this has only a 1 unit impact on the inflation. As another example, assume your league had $1500 of value in the draft but $1800 of salary to spend. This is a 20% inflation factor and means that a player with a True Value of $20 would have a Draft Value of $24. If the first 10 players ALL went for $10 more than their Draft Values (using Barry Bonds from above he would go for $64) and your league spent $520 on $420 of Draft Value (which therefore was $350 of True Value) you would have $1150 of True Value left with $1280 of salary to be spent. Thus, that $20 player would now be listed as $22. Despite massive overbidding you have only changed the draft value by $2 at the high end players. Quite simply, stick to your original assessment and buy players when they become bargains.
If players do go for significantly more than their draft value in the early rounds, that would actually drive down the inflation factor (as noted above). Thus, if you have been hanging back you will start to miss some good players who will look like bargains the next day. You will end up reducing the amount that you will spend because the other owners "overspent" ie, after each round the Draft Values of each player will decrease a little bit and you will keep hanging back waiting for bargains off of this lower number. This means you are likely to leave the draft with unspent salary units and not as good a team as you could have had. You may end up with 23 players that are all $5 bargains, but if you only spend $46 and get players worth $161 you will easily be the worst team in the league..
Remember leaving the draft with 20 unspent salary units means you could have substituted a 1 unit player on your roster for a 21 unit player if you had drafted properly. Even if that $1 player ends up being worth $6 that is still $15 of stats that are lost. That is a significant hole to put yourself in to start the season. As noted in the essay on Draft Inflation , however, some of the best bargains turn out to be the very first players taken in the draft. Never forget, the object is to buy bargains WHILE using all of your salary cap.
Thus, if the use of the computer makes you hang back, when you do jump in you may be able to get anyone you want. The question is, have you waited so long that there is no one you want to take?
The second major problem with updating the draft inflation during the draft
is that it can ultimately leads to absurd answers. For example, Suppose you had
the following situation:
Owner A: $20; Needs 1 OF, 1 catcher
Owner B: $10; Needs 1 OF
Owner C: $ 3. Needs 1 catcher
Now assume the 4 players left to be drafted have the
following values:
|
Impact of Ongoing Inflation Calculations |
|||
|
Player |
True Value |
Pre Draft
Inflation |
During Draft
Inflation |
|
Outfielder 1 |
$5 |
$6 |
$15 |
|
Outfielder 2 |
$3 |
$4 |
$9 |
|
Catcher 3 |
$2 |
$2 |
$6 |
|
Catcher 4 |
$1 |
$1 |
$3 |
In this table, True Value is the value of the player before calculating any draft inflation. Pre Draft Inflation is the draft inflated values of players after the protected lists are submitted but before the draft begins. During Draft Inflation is the rolling calculation of a player's value based on the salary cap remaining and players remaining. Also, for this illustration assume that all owners value the 4 players in the same order. For the predraft inflation we assumed a 20% inflation rate. For the During draft inflation, we have 33 draft dollars chasing $11 of true value. Thus, the computer assigns a value to each player which is 3 times higher than his True Value.
Obviously, Owner A can have any two players he wants and the prices will be determined by who introduces first. If Owner A introduces first, he would introduce catcher 3 for $3 thus assuring that player and assuring that he only overpays by $1. Obviously in this scenario Owner C takes catcher 4 for $1. Owner B then introduces outfielder 1 for $1. While we have assumed that the Owners all value the players in the same order, the Owners themselves cannot make such an assumption. Thus, if Owner A did not want outfielder 1, then Owner B could get him for $1. Thus he would only introduce him for $1 not $10 since Owner A could still take him for $11. Thus bidding $10, Owner B risks paying $9 more than he has to. A potentially major mistake in carry-over or perpetual leagues. The same reasoning dictates that Owner A will introduce the catcher first rather than the outfielder.
If Owner B bids first he will introduce outfielder 1 at $1 and he should go to Owner A for $10 or $11.
If Owner C introduces, catcher 3 should go to Owner A for $3 or $4.
The net result here is that no matter who introduces, the players should go
for the following amounts:
Outfielder 1: $10
Outfielder 2: $ 1
Catcher 3: $ 3
Catcher 4 $1
Thus, all of those players go for values that are closer to the inflated values we projected before the draft began than to the values that the computer generates during the draft!
Also, for additional dangers of starting with the wrong assumption of your league's salary allocation between offense and pitching (which also leads to absurd answers as the draft moves on) see our essay on Phantom Inflation . This essay basically discusses the impact on your draft values (especially if you are re-calculating during the draft) when the amount that the league ends up allocating to hitting and pitching varies from the way you programmed the computer to calculate the values.
Finally, a perceived use of the computer is to see where each team in your league stands in the various categories in the middle of the draft. This is a completely useless exercise. What if you have 10 hitters and the person next to you has 5? What is the validity of looking at the projected performances? You are trying to compare apples and oranges and it just doesn't work. What if you know that one team is running away with homers. Does that mean he will stop bidding on power hitters or will he try to accumulate even more to be able to deal them? If you do not know the computer surely does not and it cannot give you any information that does anything other than create a false sense of security or panic.
Time
Time
Time
The bottom line (as you should have figured out from the above) is that using a
computer will actually take more time from keeping an eye on the important
things in the draft than using pencil and paper will. However, if given enough
time to input the data, the computer can be a good organizing tool. The key is
not to give the person using the computer time. If someone kept asking you to
slow down so they could look something up in all the books they brought in, you
would laugh at them. Why then would you slow down the proceedings so the person
with the computer can catch up? He brought it in to beat you, to get an
advantage. Turn it against him by moving too quickly for him.
The simple thing to do is to outlaw them. Many leagues have since the "richer" owners can afford laptops while the others cannot. Whle this is a valid argument, the computer hurts more people than it helps. I would let anyone who wants to bring one in do so and then move quickly and watch them have the worst draft ever.
We would certainly like to hear your thoughts on this subject as well as any "war stories" you can relate about computers in the draft.
Return to Strategic Essay List
Return to PPI Home Page
Email questions to PPI or email
support@stallvalue.com
Copyright (c) 2003, Player Projections, Inc. All rights reserved.