The Instigator
mongeese
Pro (for)
Winning
12 Points
The Contender
Ore_Ele
Con (against)
Losing
11 Points

On Debate.org debates, there is, on balance, a net Contender advantage.

Do you like this debate?NoYes+3
Add this debate to Google Add this debate to Delicious Add this debate to FaceBook Add this debate to Digg  
Post Voting Period
The voting period for this debate has ended.
after 6 votes the winner is...
mongeese
Voting Style: Open Point System: 7 Point
Started: 11/2/2011 Category: Miscellaneous
Updated: 5 years ago Status: Post Voting Period
Viewed: 5,390 times Debate No: 19103
Debate Rounds (4)
Comments (76)
Votes (6)

 

mongeese

Pro

Essentially, this debate is about whether or not the Contender has an artificially higher chance of winning any given debate because he or she is the Contender. "On balance" means that not every debate needs to have a Contender advantage, and "net" means that the Instigator doesn't necessarily have no advantage.

This assumes that neither debater is acting naive or forfeiting any rounds, and that the Contender had the option to decline the Instigator's challenge with no repercussions OR the debate's resolution and rules were chosen by an unbiased third party, with both debaters agreeing to them, as most debates are.

The first round will be for acceptance only. If any clarifications need to be made, they should be requested in the Comments section.

The burden of proof is not absolute; Pro must demonstrate that the resolution overwhelmingly likely, either at least 95% probable statistically or through other methods.

Good luck to whoever accepts this debate.
Ore_Ele

Con

I thank my opponent, Mongeese, for creating this interesting opportunity. This looks to be a much better debate than the one I had against Truthseeker about the "last word advantage."

I wish to only re-iterate the BoP, as stated by my opponent. He said, "The burden of proof is not absolute; Pro must demonstrate that the resolution overwhelmingly likely, either at least 95% probable statistically or through other methods."

This means, that my main objective is to cast doubt that their is an net contender advantage, and that the advantage is held by some other feature than the instigator/contender position.

As such I will let my opponent begin.
Debate Round No. 1
mongeese

Pro

Thank you, Ore_Ele, for accepting this debate.

I will support my position from two fronts: general trends and specific examples.

I. General Trends

Of the many thousands of debates on Debate.org, 13,285 have not ended in a draw. Of those, the Instigator has won 4,578 of them and the Contender has won 8,680 of them. In performing a statistical test, our null hypothesis would be that the Instigator would win 0.5 times per debate. However, when a one-sided one-proportion z-test is performed on this data, one gets a z-value of -35.823, which means that the probability that this imbalance is due to random chance alone is 0.

I will acknowledge, however, that not all of the assumptions for a one-proportion z-test are met, although in this case I see no reason for the statistical conclusion to be any less valid.

The only significant flaw here is that while it is proven that the Contender wins significantly more often than the Instigator, because Debate.org is not a controlled experiment, it is conceivable that the characteristics of frequent Instigators and frequent Contenders are significantly different. Therefore, specific debates will be necessary to affirm the resolution, as shown in the next point.

II. Specific Examples

a. Simple Falsehoods

Often enough, the Contender makes a claim that, while apparently true, can be easily refuted by the Instigator if he were given a chance to refute. This allows the Contender to make some claims that the Instigator may not challenge; the Instigator has no such opportunity. Therefore, the Contender has an advantage in the debate. Examples:

Debate 1: http://www.debate.org...

The claim: "Plus, there are exceptions to everything. An example is that fireworks are illegal in the state of NY, but on the 4th of July, there is a special fireworks display (and fireworks can be dangerous... and yes, they do this on public property) completely within legal parameters."

The Contender tried to use an exception to the illegal status of fireworks in New York to claim that flag burning itself should not be illegal. At first glance, this makes sense, but the Instigator never got the chance to point out that just because there is an exception to a law does not mean, in the Contender's own wording, that the action is no longer illegal, making the entire point moot.

Debate 2: http://www.debate.org...

The claim: "Once again, Con is trying to manipulate definitions -- he notes that Dictionary.com's definition of omnipotent is regarded as very great power, when in fact I have pointed out in R1 that Dictionary.com's definition reads having very great or UNLIMITED (a.k.a. ALL) authority or power. This is not debatable -- a quick look at the source will reveal which one of us is being truthful and which one of us is omitting relevant information to further their position. My definition is correct, stands, and is solely instrumental in negating the resolution even without considering my other arguments."

This claim also seems to make sense in the context of the debate, until one realizes that the definition reads "great OR unlimited" instead of "great AND unlimited." The Contender's entire case rested on the idea that God's omnipotence was necessarily unlimited power, so an alternative definition of omnipotence that did not involve unlimited ability, as presented in the "or" clause, would refute her completely. The Contender naturally neglected to point out this crippling difference, skipping over it as if it were insignificant, and the Instigator never had the opportunity. The Contender was therefore able to win the debate with a flat-out incorrect argument, but none of the voters got to know that.

b. New Arguments

Often, the Contender is able to make a new argument in the final round of the debate that ought to have been introduced and easily could have been introduced much earlier in the debate. This gives the Instigator no opportunity to refute at all, even if the argument is objectionable. This is especially egregious if the argument is itself a turn used to turn an argument made by the Instigator rounds before against the Instigator, seemingly ruining the Instigator without ever standing up to criticism. Examples:

Debate 1: http://www.debate.org...

The claim: "He says that the cheaper system is the better system, and I have proven that the 360 is actually LESS expensive than the Wii."

The Contender made this argument in the very last round of the debate; the Instigator had been pushing the price point since the very first round of the debate. Had the Contender been the Instigator instead, she would either have to give the original Instigator a chance for refutation or not make the point at all.

c. Trailing Argument

Sometimes, the back-and-forth on a debate is so intense that both debaters continually try to one-up each other's responses in an attempt to disprove the other. In such an event, the point is often completely unresolved by the end of the debate. As such, in the last round, the Instigator refutes the Contender's penultimate point, but the Contender may refute the Instigator's final point, with his own point left unchallenged whether it is right or wrong. Examples:

Debate 1: http://www.debate.org...

The claim: "I think my opponent is just wrong about the cards effects. He argues that chain energy did not do the effect that the card said it should. I definitely think it did, the card drained 500 life points and then by binding them with a ring. Moments later, after the ring drains the point, they are let go and they show up in time for Dark Bakura to summon some more monsters."

This claim was regarding the effect of a Duel Monsters card in a Yu-Gi-Oh! episode. The Contender claimed that the card may have drained life points; however, any Yu-Gi-Oh! fan would be able to point out that the victims of the card were not screaming in pain, and therefore were not losing any life points, disproving the Contender's speculation. Sadly, the Instigator could not point this out, allowing the judges of the debate to be falsely convinced that the Contender was right on a crucial point when he was wrong.

Debate 2: http://www.debate.org...

The claim: "For another, the Patronuses can't fly. It would have no use in the mid-air battle. My opponent argues that the stag would run into Midna. Run, not fly."

It was completely necessary for the Contender to make this point in Round 5, because through the debate, it wasn't determined that Midna would attack Harry Potter until Round 4, leaving the Instigator to introduce the Patronus defense in Round 5. A voter would read the claim that Patronuses cannot fly and accept it as a valid refutation. However, had the Instigator and Contender switched places, the Instigator would have been able to refute this claim through evidence, such as this picture that demonstrates a Patronus not flying, but running on air, which is satisfactory to the conditions that he originally described: http://harrypotter.wikia.com...

All three of these contentions can probably be summarized in a few sentences: because the Contender gets one round of argumentation to which the Instigator has no opportunity to reply, he or she may make statements that only need to sound plausible in order to persuade a voter without risk of being countered in any way by the Instigator. Had the Instigator and Contender switched places in such a debate, the Instigator would have had a chance to refute the claims made by the Contender, significantly altering how the voting on the debate would turn out.

With that, I leave the floor to my opponent. Good luck, Ore_Ele.
Ore_Ele

Con

I thank my opponent again for this meta-debate (a debate of debates). For this, I will focus mostly on general trends, since the resolution is "on balance" and "net" advantages, but I will also make sure to address the specific examples.

Remember, Pro must show,
1)There is, on balance, a net advantage, for the contender.
2)That the advantage is due to them being the contender and not some other factor (such has being a higher skilled debater).

General Trends

This alone would appear to be a very damning piece of evidence. It is not a sample size, it is all debates, so one cannot argue that it is not an accurate representation of the whole. However, there are some things that are left out. The first is looking into alternative hypothesizes to explain the difference. Should other hypothesizes, logical and statistical tests must be done to find out which is accurate. So here is my alternative hypothesizes. The difference is caused, not by a natural bias in the system, but by the different average skills of the debaters.

For that, I'd like to give my first piece of evidence. Different average ELO scores. Of those 13,285 debates, the instigator has a weighted average ELO score of 2,230.95 and the contender has a weighted average ELO score of 2,847.49 [1, you can see the formula as well as the result]. This indicates that based on the skill difference, an average of .5 wins per debate for the instigator would not be expected, and so the stats that were run in my opponent's R2 no longer prove anything. As such, this evidence does not prove, nor disprove either theory, for that, please continue reading.

We should also remember that my opponent said "This assumes that neither debater is acting naive or forfeiting any rounds." Now, I don't think it is a stretch to claim that many debates have instigator forfeits, and I hope that wasting time and space gathering evidence will not be needed for that assumption (but if needed, I'll gladly prove it), that said, I believe it would be better to look at a selected population, where such issues (like forfeiting) is less likely to occur, and so have less of an effect.

For this, I looked at 2 different groups (all based on ELO rankings, which the top 150 have been made public). The first that we'll look at is "3k v 3k," and "4k+ v 4k+" (or top 29).

1)3k v 3k

What this is, is looking at all the debates of members with ELO rankings around 3,000 (+- 200) verse other members of the same ELO ranking. All in all, there are 32 debates which fall in this category, 9 of them since 2010 (when the voter tab became public and vote bombing because less common, though not extinct) [2]. With these debates, the ELO difference has been mostly negated, we find that the average instigator is within 100 ELO points of the average contender. We also find that the instigator has won 16 of the 32 debates (a dead split). Should there be an "on balance" contender advantage, this would not be expected, and must have some explanation as to why this differs from the theory.

I would also like to take this opportunity to show the significance of the Null hypothesis (meaning that the 50% win is not a fluke), much like my opponent did in his round. We find a z-value of 0, and we can conclude with 100% certainty that this is not random [3].

2)4k+ v 4k+

Much like the 3k v 3k, this was done on all debates of which members with ELOs of 4,000 or above debated other members of 4,000 and above. Here, we actually see somewhat of a halfway between the 3k v 3k, and the total debates. Among these members, there were 121 debates, which ended in a 55 – 66 record for the instigator (or 45.45% wins). However, like with the total debates, we do see a difference of average ELO rankings between the instigators and contenders. This time of about 300 points (5318.12 v 5608.93). Like my theory would predict, the higher ELO score would give them an advantage, however less significant do to the scores being closer.

Prediction abilities

This actually starts to fall on a rather linear relationship between instigator ELO scores vs contender ELO scores and instigator win%. Simply the deviation from 50% is (I-C)/2I (where I = Instigator ELO and C = contender ELO). With the standard group, we find that the large difference in ELO would estimate that the Instigator wins 36.18% of the time (real numbers are 34.46%), the difference between real and predicted numbers is easily explained through the forfeits which are bias against instigators. When looking at the 4k+ debates, it predicts a 47.26% win for the instigators (real numbers are 45.45%), this is well within the margin of error and is not rejected by any significance level [4]. With the 3k debates, this would predict that a 51.08% win for the instigators (real numbers are 50.00%). This, too, is well within the margin of error so that it passes all significance levels.

Simplicity

For the prediction, we can see that it is pretty darn close. Now, of course, it is entirely possible that I tried to force some equation to fit these various tests (in which case, one could not claim prediction, since that would be circular). However, the simplicity of (I-C)/2I would indicate no forcing was involved. When you see forcing, you'll often see a number of coefficients whose only purpose is to make it work. Here we have no coefficients at all.

Specific Examples

My opponent goes on and lists 5 specific examples which he believes shows a particular advantage to the contender. Oddly enough, on these, every opposing debater has a higher ranked ELO, and in one example, despite the "contender advantage" the instigator still won. Now, I'm not going to get into debating every little feature of every debate that my opponent feels the instigator should have won. Even with 800,000 characters, that would never be enough. However, while there are some advantages to being the contender, there are also advantages to being instigator, such as setting the resolution and definitions to best suit your arguments (can be very effective if your opponent does not see how the definitions aid you). Cherry picked examples cannot possibly show that, on balance, the contender has the net advantage.

Thank you

[1] http://www.debate.org...
[2] http://www.debate.org...
[3] http://www.wolframalpha.com...
[4] http://www.wolframalpha.com...
Debate Round No. 2
mongeese

Pro

Thank you, Ore_Ele, for your response.

I. General Trends

My opponent proposes an alternate hypothesis to the Contender advantage, claiming that instead the Instigator is on average less skilled than the Contender. To prove this, he uses ELO scores themselves based on the results of the debate. This reasoning is somewhat circular, as it uses the results of the debates to decide the skill of the debaters, leaving nowhere to eliminate any theoretical Contender advantage. Should a Contender advantage exist, the supposed difference of skill in the debaters would actually be caused by this advantage; frequent Contenders would win debates more often, thus appearing more skilled under my opponent's system. Essentially, this leaves us nowhere; if a Contender advantage exists, then it will explain the ELO disparity, and if it doesn't, then it won't. Therefore, all of my opponent's conclusions drawn from ELO statistics are biased and useless, presupposing that the resolution is false to prove that the resolution is false.

Look at my opponent's statistics regarding the 3k group. If a Contender advantage exists, then the frequent Instigators in this group are actually more skilled than they appear, and frequent Contenders are actually less skilled. Therefore, an evenly split win ratio is to be expected. This assumes, of course, that the ELO rank is perfectly reliable, and that nobody reverses their side; if people are shifting around, then there could also be an additional Contender advantage despite all attempts to quell it - and there is.

My opponent is correct in that many debates have Instigator forfeits; however, many debates have Contender forfeits, as well. In fact, many of the debates that my opponent uses in an attempt to refute the Contender advantage, in his 3k group, are Contender forfeits (6590, 3546, 10565), partly-forfeits (4095, 7875, 14417), naivety (3563, 5825, 16751), or the result of vote-bombing and only vote-bombing (8375, 8543). Meanwhile, only one Instigator victory (8340) should be discarded. Since most Instigator victories don't qualify for this debate, we are instead left with 5 victories out of 20 debates, a dismal 0.25 win ratio that a one-proportion z-test says could only occur with 0.0127 probability due to random chance alone. With our established 95% threshold, we must reject the null hypothesis and accept that there is a Contender advantage.

I would apply this same reasoning to his 4k group, but he does not link to the debates, and there are way too many debates for me to want to properly analyze all of them. I believe that the evidence on the 3k group should suffice to expose the true trend. My opponent's additional analyses of his statistics are irrelevant, as I've shown that they are based on faulty assumptions and poorly handled data.

My opponent also doesn't seem to properly understand how proportion statistics work. When analyzing his 3k and 4k group data, he chose to use a z-test instead of a one-proportion z-test, making up a standard deviation of 1, which is a clear indicator that he is using the wrong test; we are dealing with successes and failures, which means we must use proportion statistics. If one replaces the standard deviation with another value, like 0.1, then the null hypothesis is rejected at the 5% significance level that we've chosen. If we apply a one-proportion z-test to his data instead, as we should, we get instead a p-value of 0.130. If we further assume that we really should expect the Instigator to win as often as the Contender because of the Contender bias present in the ELO system, we get a p-value of 0.042, which passes the necessary threshold; we therefore hold that there is a Contender advantage.

To conclude my point, my opponent's data collection, assumptions, and statistical tests are all faulty, and when we correct for these failures, we find evidence not for neutrality but for a Contender advantage, turning this entire point.

II. Specific Examples

It is naturally unsurprising that the Contenders have high ELO scores if one assumes that there is a Contender advantage and the Contender wins artificially more often than the Instigator; it goes back to the circular reasoning behind using ELO scores that I identified before. The one Instigator victory in the group was a close one, by 2 points, as an artificial advantage does not guarantee victory; it need only appear and be identified for proof of its existence.

My opponent makes no attempt to refute the apparent Contender advantage gained from having the last word, claiming that it would take too many characters. I identified five claims made, then explained why they are examples of the Contender advantage with fewer than 3,000 characters, repeating myself considerably often; he should require no more than that for refutations.

Instead, he supposes some vague Instigator advantages, but these do not appear to be advantages at all. The Instigator may set the resolution and definitions, but the Contender may negotiate both before accepting the debate, sometimes even rejecting the definitions after they have been chosen by the Instigator, appearing to be instead a disadvantage instead of an advantage for the Instigator (10398).

My opponent also claims that the Instigator can essentially trick the Contender with definitions that are biased towards the Instigator without appearing to be biased towards the Instigator, or the Contender can simply misunderstand. I can only recall two debates in which this appeared to occur in any form (8426, 8527). In the first, the voters rather overwhelmingly decided that the trickery was poor conduct and voted for the Contender instead. In the second, the Contender misunderstood the definitions by his own naivety.

Sometimes, the Instigator can even shoot himself in the foot with definitions that make his own case much more difficult (8088, 10752), but this should be attributed not to any inherent advantages or disadvantages but instead to the naivety of the Instigator. Likewise, a Contender who accepts a debate in which the definitions are stacked against him without using his power to negotiate is equally naive.

Additionally, the Instigator is traditionally charged with the burden of proof. This wouldn't be much of a problem if it weren't for the fact that the Instigator also runs into problems transferring the burden even in the first round, before acceptance (8601), giving the Contender even more advantage.

My opponent accuses me of cherry-picking my examples, but there is no more effective way to demonstrate the Contender advantage than to identify examples of it in action. It could legitimately be cherry-picking if I consciously ignored any examples of Instigator advantage to show that only a Contender advantage exists, but I don't believe that such an advantage exists, and my opponent likewise hasn't come up with any concrete examples, either. If my opponent believes that there are any debates that demonstrate an Instigator advantage that I have skipped over, then I suggest he post them. As it stands, we only have examples of Contender advantage, so if we assume that sometimes a debate has a Contender advantage and other times there is no advantage, then when everything is taken as a whole, there is overall a slight Contender advantage. It's simple math; the average of a thousand zeros and a few negative numbers will come out negative.

In conclusion, not only do the statistics clearly indicate that a Contender advantage exists despite my opponent's best efforts, but we also have solid examples of the Contender advantage in action.

Good luck, Ore_Ele.

[Sources: All debates sourced are referred to by their debate number. All z-test statistics came from Wolfram Alpha. All one-proportion z-test statistics were one-sided and came from a TI-84 graphing calculator, using the data referenced and a p0 value of 0.5 unless otherwise specified.]
Ore_Ele

Con

I would like to thank my opponent for his round.

I) General Trends

I must question his removing of certain debates from the 3k v 3k. Several issues arise, in both his method for choosing which debates to cut, and how he counted. For one, he said that debate 3563 [1], should be thrown out of the Instigator win column because of "naivety." The problem with this is that debate 3563 is not an Instigator win to begin with, it is a contender win. Therefore, you cannot discard it from the instigator win category. Another issue is the discarding of "partial forfeits" such as 4095 [2]. We can see that the contender did forfeit 1 round, of which the instigator also skipped his round out of courtesy. Through the other rounds, a quality debate was able to take place, and as such, voting should be considered legit.

Now, my opponent brings up concerns of vote bombing. It should be known that prior to 2010 (I think it was Feb 2010 that it was opened, but not sure, I just remember that it was shortly after I joined), the voting tab was closed, so vote bombing likely occurred, a lot. And there is virtually no way to prove it. The only ways are when people admit it, were caught in a lie about their vote (see debate 8543 or the Danielle one that Mongeese has in his photos), or the IP addresses are matched up against voting records (we've only been able to dig into that recently). As such, it is impossible to accurately say that one debate was VB'd and another one wasn't. This is why I separated out the pre-2010 and post-2010 debates. While vote bombing did not completely stop, it certainly was greatly reduced (and reduced again with RFD's).

If we exclude the debates that have a high chance of vote bombing, we still find that the instigator won 4 – 3 (if we toss out 10565, which was the only complete forfeit). Of course, then we find our sampling to be very small.

If we go to the 4k+, we find 121 debates (There was not enough room to make a source link for that many, since each debate takes up about 90 characters, and the limit is 8,000), now for these, if we limit them to the open tab era, we find 29 of them [3]. With these, we see 12 instigator wins (41.4%) and the ELO difference (5199 – 5798) would estimate a 44.2% win ratio (2.8% variance).

For more evidence, I've gone through (several times now) and randomly selected a handful of debates to test my ELO win predictor. In the group that I did, I had the following debates [5]. We had 60 debates. The average Instigator ELO was 2255.52, the average contender was 2885.08. This yields a predicted instigator win rate of 36.04% (which would be 22 wins), we find that 23 of the 60 were instigator wins, a 38.33%. The prediction was off by only 1 win, and it under-estimated (many of the past samplings, like the 3k v 3k and 4k+ it was slightly over estimating).

However, I realize that I cannot prove that this is a true random sampling, as such, I've presented a request in the comments that anyone (though I specifically asked F-16, since he was the last to post a comment) please offer up 50 - 80 random numbers from 1 - 13591 so that I can do this test for them all.

I.a) Questioning the ELO Gods?!

Okay, so referencing the ELO Gods is just a joke, however, my opponent does bring up a legitimate point. ELO is based on a person's wins, so it presents the risk of circular reasoning. However, I'd like to point out an issue. Any measure of skill is going to be dependent upon the results (in this case, one's debate skills are dependent upon the debate results), ELO has proven to be a very accurate measuring tool in countless applications [4], from Chess, to college football, to Battlenet. While, of course, any system is going to be imperfect, ELO has shown itself to be one of the most effective measuring tools available. And it's accuracy is only enhanced by the fact that the simple prediction tool I made in the last round was still effective in the post-2010 4k+ (off by 2.8%).

II) Specific examples.

Moving on to the Instigator advantages, there are of course limits. Just like with contender advantages. The contender has the advantage of the last word, but if they abuse this and post a bunch of new arguments (19025), that will actually work against them. Likewise, the same can happen with definitions (10398), if the instigator presents wiki definitions and the contender presents medical definitions, the advantage can be negated. It is the simple that it you push any advantage too far, it will swing back and bite you. Some cases where definitions are presented and hold for true advantage are 10982 and 15819. We can also see this used in many religious debates (where the instigator defines "God"), such as 17753, 17901, etc. Of course, failure to define allows definitions to be passed to the contender (as seen on 16910, and so many semantic victories), thus passing on the advantage. But the instigator has the first pick and first right to that advantage.

My opponent tries to negate these advantages by saying that if someone walks into a trap, it is their own naivety. While that is so, one could place the same naivety upon any instigator. Everyone that starts a debate knows that their opponent will have the last word, therefore, if they are not prepared for that, we can claim that it is their own naivety. But just because they are being "na�ve" (which is a subjective value anyway) does not mean that there is no advantage.

As such, there are many advantages and powers that both sides have (from last word, to setting definitions, to setting the BoP, to negotiating, etc), but simply pointing out one advantage over another does not show any kind of NET, ON BALANCE, ADVANTAGE.

[1] http://www.debate.org...
[2] http://www.debate.org...
[3] 10982, 11353, 11937, 12249, 12301, 12409, 12486, 12517, 12616, 13115, 13130, 13260, 13307, 13320, 13579, 13988, 14083, 14284, 14526, 14642, 15081, 15819, 16769, 16792, 16826, 16947, 17029, 17214, 17925.
[4] http://en.wikipedia.org...
[5] 250, 349, 526, 779, 1033, 1309, 1570, 1755, 1840, 1994, 2101, 2246, 2435, 2595, 2826, 2986, 3182, 3428, 3669, 3922, 4097, 4273, 4454, 4655, 4852, 5125, 5438, 5846, 6265, 6617, 6954, 7304, 7708, 8068, 8302, 8574, 8932, 9135, 9323, 9594, 9896, 10245, 10633, 10954, 11453, 11540, 11683, 11888, 12163, 12382, 12393, 12646, 12747, 12844, 13043, 13220, 13371, 13499, 13550, 13583.
Debate Round No. 3
mongeese

Pro

Thank you, Ore_Ele, for your response.

I) General Trends

I admit, I did mistakenly regard Debate 3563 as an Instigator victory instead of a Contender victory. I also realize that I made the same error on 7875. This would mean the Instigator won 7 of 20 debates, which still leaves a p-value of 0.090, rather low, and supports my arguments in conjunction with other tests to demonstrate a statistically significant Contender advantage.

Regarding the partial forfeit, recall that one of the qualifiers for this debate is a complete lack of forfeits, not just ones that my opponent would consider significant, so we must still disqualify 4095. Debate 14417 was won only because of the single forfeited round, by a single conduct vote.

As for vote-bombing, the only two debates that were vote-bombed to Instigator victories were those by untitled_entity, who actually has an apparent pattern of winning via rather undeniable vote-bomb (8543, 8421) and was identified by my opponent as a vote-bomber, making her debates easy to remove. My two removals were clearly justified, and my opponent has no ground to dispute them. Looking at only post-tab debates leaves too small of a sample size.

My opponent then analyzes other random sets of debates, but every time, the Instigator wins statistically significantly less often than the Contender. He justifies this with the difference in ELO scores, but this still relies on trusting the ELO scores to be reasonably accurate in this case, which is circular reasoning.

My opponent is correct in ELO's success in various fields, but DDO's debates are fundamentally different because the Instigator is not selected randomly, and some people have a greater inclination to instigate a debate than others, so any Contender bias would not be evenly distributed among the population. In chess, for example, there is an advantage to moving first [1], but because in tournaments, everyone goes first as often as they go last thanks to random determination, this is no problem. If, however, some people play as White more often and others play as Black more often (which wouldn't reasonably happen in chess, but does in DDO, hence why my opponent's comparison is invalid), those who play White more often will also win more often, and thus have higher ELO scores, even if everyone in the chess league was at the exact same level of chess skill - my opponent would tell you, however, that this is because those who play a White more often are also inherently more skilled, even though this is false by premise. Essentially, the ELO system will almost tautologically be somewhat accurate in predicting win and loss percentages in any case, whether there is bias or not, making its presence in this debate irrelevant.

To conclude this point, once one removes the circular reasoning involved, one sees that my opponent's statistics supports the concept of a Contender advantage very strongly.

II. Specific examples

My opponent claims that Debate 19025 is an example of a Contender advantage pushed too far, but I see nothing in the RFDs or Comments section concerning any last-round abuse. He claims to have four debates that demonstrate Instigator advantage:

10982: The Contender stated in the Comments section, "I think I truly messed it up by not defining 'Easter Bunny' more concretely in the first round." This was not an advantage from being Instigator, but from having an opponent that failed to define things, which happens to the Contender just as often as, if not more often than, the Instigator. I have quite a few semantics debates of my own, Contender in all of them, and my opponent acknowledges their plentitude of existence, so I see no need to point them out.

15819: I don't see the inherent advantage here. The Contender challenged the Instigator's definition as being "entirely inconsistent," and turned the debate into one about the acceptability of the Instigator's definition. It became the center of the entire debate. The Contender controlled the course of the debate in his acceptance round; he only went wrong in that he subsequently lost the argument that he thought he would win, and therefore had to debate with the Instigator's definition, which cost him the debate. A debate in which the Contender may identify which points will be debated and with what priority before even having to accept cannot possibly be interpreted as having an Instigator advantage.

17753: The Instigator never defined "God," despite my opponent's false claim. He did, however, define "omnipotence" in the second round. However, had the Contender disagreed with the definitions, he could easily have challenged them. There was no Instigator advantage; the two debaters were on equal footing in that regard.

17901: The Instigator defined "God," yes, but the Contender accepted the definition completely. Had the debate been reversed, the Contender would have chosen the same definition. If he had any problem with the definition, he would have proposed a change beforehand. There was no advantage here.

Sadly, my opponent took no time to clarify in any real depth how these debates show an Instigator advantage as I did for my examples of Contender advantage, so it's considerably likely that he will point out that he meant something completely different to what I refuted. He would then clarify, and I would be powerless to respond. If such an event should occur, it would be a Trailing Argument Contender Advantage.

My opponent then claims that every Instigator could be identified as naive for instigating a debate in which the Contender has an inherent advantage. However, any Contender advantage cannot be prevented by the Instigator, as it takes place in the last round without forewarning, so it is essentially impossible to prepare for the Contender advantage except by not instigating debates at all, in which case no one wins. My opponent did not explain how it could be prevented; I could try to predict all of the ways in which he claims they could be prevented, as he will probably explain at my request during the next round, but it would be character-wise, putting me at a new disadvantage as it draws space from my other arguments to refute points that may never be brought up anyway, and even then, I could miss the mark completely, making the entire attempt a waste and giving my opponent a huge Contender advantage.

The two scenarios of advantage are also completely different. It's the difference between buying an average house that is later bombed by an invading army, and buying a house that is clearly marked and scheduled for a demolition tomorrow. One instance is unpreventable except by complete avoidance of ever buying a house; the other can easily be foreseen and thus avoided by either cancelling the demolition before purchase or not buying that particular house.

Of course, even with a broad interpretation of naivety, there is an exception: tournament debates. Neither debater chooses to be Instigator, and thus neither debater can be considered naive for that; however, we still end up with an Instigator and a Contender, and thus a Contender advantage such as those I outlined with five examples.

My opponent then adds that naivety does not eliminate an Instigator advantage. This is true, but since a qualifier for this debate was a lack of naivety, all advantage by naivety, which seems to be all true Instigator advantage, is irrelevant to this debate.

To conclude this point, I have five clear examples of undeniable Contender advantage, but my opponent offers only vague descriptions of theoretical examples or misinterpretations of reality.

In conclusion, as proven by both statistical tests of significance and specific analysis, the Contender advantage clearly exists.

Thank you, Ore_Ele, for this entertaining meta-debate. Good luck in your final round; try not to stretch your advantage too much. ;)

1. http://en.wikipedia.org...
Ore_Ele

Con

I thank my opponent for a most interesting debate about the advantages and disadvantages for each side of the debate. I will address my opponent's arguments, but I will focus mostly on summarizing.

First things first. F-16, Roy, and Danielle, each provided a list of random debates to test my formula. I'm happy to say that it still remains fairly accurate. Of their 215 listed debates, 158 have winners, 64 of them were instigators. The average ELO of the instigator was 2487.8 and the average ELO of the contender was 2863.4, this would predict a 42.3% win rate for the instigators. What we saw was 40.5% (1.8% off, not bad at all). Of the three, Roy turned out to be the closest, being off by less than 0.3%.

However, this is really moot, since my opponent claimed that these ELO score are meaningless, and that they don't take consideration of the "contender advantage." As I pointed out in a previous round, ELO is used for many many things (this cannot be viewed as a new argument, because I brought it up last round). My opponent only addresses Chess, however, ELO is used for things from Starcraft to College Football. And we can certainly see that there are advantages to different races in Starcraft, and different stadiums have different degrees of "home field advantage." So in these cases, the advantage is not evened out (as it is in Chess), yet ELO still remains highly accurate in those. There is no reason to believe that it would not maintain its high accuracy here for Debates on DDO.

As for the specific examples, my opponent seems to think that simply because one can challenge things before accepting a debate, that the advantage is neutralized. This is clearly false, like I said previously, the instigator can say "no" and not change things.

My opponent then brings up, probably the most damning piece against himself. He says, "it's considerably likely that he will point out that he meant something completely different to what I refuted. He would then clarify, and I would be powerless to respond. If such an event should occur, it would be a Trailing Argument Contender Advantage." Then follows with, "any Contender advantage cannot be prevented by the Instigator, as it takes place in the last round without forewarning, so it is essentially impossible to prepare for the Contender advantage except by not instigating debates at all."

In one breath, my opponent completely defeats my advantage, because if I take advantage of it, I will lose conduct and those arguments with be dropped by most voters (for argument abuse), yet in the next, he has he [as instigator] is powerless to stop the advantage. But he just did the paragraph before. My opponent has negated himself.

In summary.

My opponent needed to show that there is, on balance, a net contender advantage. What has been shown through this debate is that the contender advantage can be stopped (as my opponent pre-empted and stopped my own last word advantage). ELO is an accurate measure of skill, even in cases where there may be an advantage (such as Starcraft and Football). Skill plays a far greater role in determining a winner of a debate, and even though the raw data seems to show that there is a contender advantage, it is actually because there is an imbalance of skill between instigators and contenders, not some inherent advantage.

But mostly, since my opponent refuted his own argument, even if you don't buy the ELO argument (hint: you shouldn't), my opponent has not met his BoP (because he refuted himself).

Thank you.
Debate Round No. 4
76 comments have been posted on this debate. Showing 1 through 10 records.
Posted by mongeese 5 years ago
mongeese
The debate wasn't created to complain about anything, just to point out that the Contender advantage exists. I picked whichever debates I knew demonstrated the most Contender advantage, and most of them happened to be yours. You have been the Contender of more debates on DDO than anyone else, so it's hardly surprising that this would happen.
Posted by Danielle 5 years ago
Danielle
Roy said that the Instigator has the advantage of selecting the resolution, and this should be an advantage because the Instigator can pick a subject he knows well. Meanwhile, the Contender is more likely to have to research it. I think that's BS because the Contender would only accept a resolution they (a) probably know pretty well already, and (b) one they think they can win. Many times people do not accept debates just because they don't think they can win, even if they disagree with the Instigator. I agree with Roy that the Instigator usually has the BOP, which is a slight disadvantage. That, however, depends on the resolution and voter. Inherently I think there is an advantage to having the last word.

That said, I have read this debate but will not vote until later when I can provide a more thorough RFD. For now I will say this: First, mongeese, as always I am more than flattered that you've created yet another debate as a platform to both complain about me and highlight your losses to me. Aww. Second, one point that the Contender didn't make in this debate (unless I missed it - that's why I wanna read this again later) that I would have made is that often the Contender does not have enough character space to present their case, because they must utilize all of their character space to negate the Instigator's claims and therefore do not have a lot of space to make their own. The Contender must keep both their rebuttal and their contentions brief because of character limits. This is one problem I've had as a Contender in the past.
Posted by mongeese 5 years ago
mongeese
Any debate can be found using it's number. If its number is XXXX, then you would use this hyperlink:
www.debate.org/debate/XXXX
Posted by 16kadams 5 years ago
16kadams
ore-ele is winnning because he is the contender ha ha!
Posted by RoyLatham 5 years ago
RoyLatham
I was commenting on the topic, not the debate.
Posted by mongeese 5 years ago
mongeese
I don't believe Ore_Ele ever made any arguments about the Contender's need to research.
Posted by RoyLatham 5 years ago
RoyLatham
I haven't made up my mind on this yet. I'm not sure there is enough evidence to reach a conclusion.

The Instigator has the advantage of selecting the resolution. That should be an advantage because the Instigator can pick a subject he knows well. The contender is more likely to have to research it.

The Contender has the advantage of (usually) the Instigator bearing the burden of proof. Many people know that new arguments made in the final round ought to be ignored, but some count them.
Posted by wiploc 5 years ago
wiploc
Pro talked about Z-things, which I didn't understand.

Also, Pro opened with a startling ratio of instigator wins to contender wins. Dramatic evidence! If he could prove that the ratio was caused by a contender advantage, big win!

But what if there is another cause? Good question, and Con asked it.

Con offered an alternative explanation, an ELO-based argument, and Pro seized on that and didn't let go. Target fixation may have cost him the debate. (I don't know what ELO is. I asked once, in the DDO forum, but nobody answered. In my experience, "ELO" stands for "Electric Light Orchestra," which may date me (I'm from the sixties).)

There was back-and-forth between sophisticated statisticians, with no winner apparent to this ignorant observer. But Pro had the burden of proof, and he voluntarily made it a really tough one. Pro wrote:

: Pro must demonstrate that the resolution overwhelmingly likely,
: either at least 95% probable statistically or through other methods.

Is it 95% likely that contender advantage is what explains the striking statistic? I don't know. Pro quite effectively showed that Con hasn't shown his alternative is the real cause of the striking statistic. But Con neglected to show that his explanation is the real cause.

And Pro undertook to prove that no alternative hypothesis is the explanation, to 95% probability. Perhaps because he focused on Con's single alternative rather than bolstering his own case or attacking alternative explanations as a group, he neglected to do so.

Winner: Con.
Posted by mongeese 5 years ago
mongeese
The irony of this debate is immense, but I think the greatest irony, from my viewpoint at least, is that I'm being defeated so far by a contention introduced in the last round that I feel more than capable of refuting, yet can't, due to the Contender advantage itself.
Posted by Maikuru 5 years ago
Maikuru
This was a very fun read. I hope to see more DDO-related debates soon.
6 votes have been placed for this debate. Showing 1 through 6 records.
Vote Placed by Contra 5 years ago
Contra
mongeeseOre_EleTied
Agreed with before the debate:Vote Checkmark--0 points
Agreed with after the debate:Vote Checkmark--0 points
Who had better conduct:Vote Checkmark--1 point
Had better spelling and grammar:--Vote Checkmark1 point
Made more convincing arguments:Vote Checkmark--3 points
Used the most reliable sources:-Vote Checkmark-2 points
Total points awarded:42 
Reasons for voting decision: Con's arguments were unnecessarily confusing, he took the Contender advantage as well in the last round.
Vote Placed by Multi_Pyrocytophage 5 years ago
Multi_Pyrocytophage
mongeeseOre_EleTied
Agreed with before the debate:--Vote Checkmark0 points
Agreed with after the debate:--Vote Checkmark0 points
Who had better conduct:Vote Checkmark--1 point
Had better spelling and grammar:--Vote Checkmark1 point
Made more convincing arguments:Vote Checkmark--3 points
Used the most reliable sources:--Vote Checkmark2 points
Total points awarded:40 
Reasons for voting decision: Con got to do a new argument in the end, which kind of proves Pro's point, and also gives him conduct.
Vote Placed by curious18 5 years ago
curious18
mongeeseOre_EleTied
Agreed with before the debate:--Vote Checkmark0 points
Agreed with after the debate:--Vote Checkmark0 points
Who had better conduct:Vote Checkmark--1 point
Had better spelling and grammar:--Vote Checkmark1 point
Made more convincing arguments:-Vote Checkmark-3 points
Used the most reliable sources:--Vote Checkmark2 points
Total points awarded:13 
Reasons for voting decision: I think Mongeese is totally right, Ore_ele totally added an argument at the end, but marking conduct against him because of it kinda proves his point. So I'm conflicted there. But his elo argument thing seemed legit. The use of debate numbers, without linking the debates, made sourcing horrible, I couldn't see any of the debates to see what was going on.
Vote Placed by 16kadams 5 years ago
16kadams
mongeeseOre_EleTied
Agreed with before the debate:Vote Checkmark--0 points
Agreed with after the debate:Vote Checkmark--0 points
Who had better conduct:--Vote Checkmark1 point
Had better spelling and grammar:--Vote Checkmark1 point
Made more convincing arguments:Vote Checkmark--3 points
Used the most reliable sources:--Vote Checkmark2 points
Total points awarded:30 
Reasons for voting decision: notice I voted pro. In my eyes he did better. Ore_ele only won becausehe had the last word, which should prove something, the last round saved him. So talk about people refuting themeselves. That is the only thing that ore_ele had won. The final round.
Vote Placed by wiploc 5 years ago
wiploc
mongeeseOre_EleTied
Agreed with before the debate:--Vote Checkmark0 points
Agreed with after the debate:--Vote Checkmark0 points
Who had better conduct:--Vote Checkmark1 point
Had better spelling and grammar:--Vote Checkmark1 point
Made more convincing arguments:-Vote Checkmark-3 points
Used the most reliable sources:--Vote Checkmark2 points
Total points awarded:03 
Reasons for voting decision: RFD in comments.
Vote Placed by Maikuru 5 years ago
Maikuru
mongeeseOre_EleTied
Agreed with before the debate:--Vote Checkmark0 points
Agreed with after the debate:--Vote Checkmark0 points
Who had better conduct:--Vote Checkmark1 point
Had better spelling and grammar:--Vote Checkmark1 point
Made more convincing arguments:-Vote Checkmark-3 points
Used the most reliable sources:--Vote Checkmark2 points
Total points awarded:03 
Reasons for voting decision: I enjoy debates concerning the dynamics of this site. Both sides discussed advantages enjoyed by instigators and contenders, though none of their examples were definitive or encompassing. Pro's preemptive strike against Con's end-round arguments was ironically self-defeating. Pro's major contention - that DDO's ELO rankings do not work within a controlled system - needed a lot more examples of abuse by users to be viable.